The recent cloning of a second estrogen receptor (ER) provided a new tool to investigate and clarify how estrogens are capable of communicating with the brain and influence gene expression and neural function. The purpose of the present study was to define the neuroanatomical organization of each receptor subtype using a side-by-side approach and to characterize the cellular population (s) expressing the ERbeta transcript in the endocrine hypothalamus using immunohistochemistry combined with in situ hybridization. Axonal transport inhibition was accomplished to cause neuropeptide accumulation into the cytoplasm and thus facilitate the detection of all positive luteinizing hormone-releasing hormone (LHRH), corticotropin-releasing factor (CRF), vasopressin (AVP), oxytocin (OT), gastrin-related peptide (GRP), and enkephalin (ENK) neurons. The genes encoding either ERalpha or -beta were expressed in numerous limbic-associated structures, and fine differences were found in terms of intensity and positive signal. Such phenomenon is best represented by the bed nucleus of the stria terminalis (BnST) and preoptic area/anterior hypothalamus, where the expression pattern of both transcripts differed across subnuclei. The novel ER was also found to be expressed quite exclusively in other hypothalamic nuclei, including the supraoptic (SON) and selective compartments (magnocellular and autonomic divisions) of the paraventricular nucleus (PVN). A high percentage of the ERbeta-expressing neurons located in the ventro- and dorsomedial PVN are of OT type; 40% of the OT-ir cells forming the medial magnocellular and ventromedial parvocellular PVN showed a clear hybridization signal for ERbeta mRNA, whereas a lower percentage (15-20%) of OT neurons were positive in the caudal parvocellular PVN and no double-labeled cells were found in the rostral PVN and other regions of the brain with the exception of the SON. Very few AVP-ir neurons expressing ERbeta transcript were found throughout the rat brain, although the medial PVN displayed some scattered double-labeled cells (<5%). Quite interestingly, the large majority of the ERbeta-positive cells in the caudal PVN were colocalized within CRF-ir perikarya. Indeed, more than 60-80% of the CRF-containing cells located in the caudolateral division of the parvocellular PVN exhibited a positive hybridization signal for ERbeta mRNA, whereas very few (<5%) neuroendocrine CRF-ir parvocellular neurons of the medial PVN expressed the gene encoding ERbeta. A small percentage of ERbeta-expressing cells in the dorsocaudal and ventromedial zones of the parvocellular PVN were also ENK positive. The ventral zone of the medial parvocellular PVN also displayed GRP-ir neurons, but no convincing hybridization signal for ERbeta was detected in this neuronal population. Finally, as previously described for the gene encoding the classic ER, LHRH neurons of both intact and colchicine-pretreated animals did not express the novel estrogen receptor. This study shows a differential pattern of expression of both receptors in the brain of intact rats and that ERbeta is expressed at various levels in distinct neuropeptidergic populations, including OT, CRF, and ENK. The influence of estrogen in mediating genomic and neuronal responses may therefore take place within these specific cellular groups in the brains of cycling as well as intact male mammals.