The P(i) concentration of mammalian cerebrospinal fluid (CSF) is about one-half that of plasma, a phenomenon also shown here in the spiny dogfish, Squalus acanthias. The objective of the present study was to characterize the possible role of the choroid plexus (CP) in determining CSF P(i) concentration. The large sheet-like fourth CP of the shark was mounted in Ussing chambers where unidirectional (33)P(i) fluxes revealed potent active transport from CSF to the blood side under short-circuited conditions. The flux ratio was 8:1 with an average transepithelial resistance of 87 ± 17.9 Ω·cm(2) and electrical potential difference of +0.9 ± 0.17 mV (CSF side positive). Active P(i) absorption from CSF was inhibited by 10 mM arsenate, 0.2 mM ouabain, Na(+)-free medium, and increasing the K(+) concentration from 5 to 100 mM. Li(+) stimulated transport twofold compared with Na(+)-free medium. Phosphonoformic acid (1 mM) had no effect on active P(i) transport. RT-PCR revealed both P(i) transporter (PiT)1 and PiT2 (SLC20 family) gene expression, but no Na(+)-P(i) cotransporter II (SLC34 family) expression, in the shark CP. PiT2 immunoreactivity was shown by immunoblot analysis and localized by immunohistochemistry in (or near) the CP apical microvillar membranes of both the shark and rat. PiT1 appeared to be localized primarily to vascular endothelial cells. Taken together, these data indicate that the CP actively removes P(i) from CSF. This process has transport properties consistent with a PiT2, Na(+)-dependent transporter that is located in the apical region of the CP epithelium.