GPER is a membrane-associated estrogen receptor of the family of G-protein coupled receptors. For breast cancer, the contribution of GPER to promoting the proliferation and migration of both carcinoma cells and cancer-associated fibroblasts (CAFs) in response to estrogen and other agonists has extensively been investigated. Intriguingly, GPER was previously found to be localized to the nucleus in one isolate of breast CAFs. Moreover, this nuclear GPER was shown to bind regulatory sequences of cancer-relevant target genes and to induce their expression. We decided to find out what induces the nuclear localization of GPER, how general this phenomenon is, and what its functional significance is. We discovered that interfering with N-linked glycosylation of GPER, either by mutation of the predicted glycosylation sites or pharmacologically with tunicamycin, drives GPER into the nucleus. Surveying a small set of CAFs from breast cancer biopsies, we found that a relatively common single nucleotide polymorphism, which results in the expression of a GPER variant with the amino acid substitution P16L, is associated with the nuclear localization of GPER. GPER with P16L fails to be glycosylated, presumably because of a conformational effect on the nearby glycosylation sites. GPER P16L is defective for membrane-associated signaling, but instead acts like an estrogen-stimulated transcription factor. In CAFs, it induces the secretion of paracrine factors that promote the migration of carcinoma cells. This raises the possibility that the GPER P16L polymorphism could be a risk factor for breast cancer.