Phosphatase of regenerating liver 3 (PRL-3) is suggested as a biomarker and therapeutic target in several cancers. It has a well-established causative role in cancer metastasis. However, little is known about its natural substrates, pathways, and biological functions, and only a few protein substrates have been suggested so far. To improve our understanding of the substrate specificity and molecular determinants of PRL-3 activity, the wild-type (WT) protein, two supposedly catalytically inactive mutants D72A and C104S, and the reported hyperactive mutant A111S were tested in vitro for substrate specificity and activity toward phosphopeptides and phosphoinositides (PIPs), their structural stability, and their ability to promote cell migration using stable HEK293 cell lines. We discovered that WT PRL-3 does not dephosphorylate the tested phosphopeptides in vitro. However, as shown by two complementary biochemical assays, PRL-3 is active toward the phosphoinositide PI(4,5)P(2). Our experimental results substantiated by molecular docking studies suggest that PRL-3 is a phosphatidylinositol 5-phosphatase. The C104S variant was shown to be not only catalytically inactive but also structurally destabilized and unable to promote cell migration, whereas WT PRL-3 promotes cell migration. The D72A mutant is structurally stable and does not dephosphorylate the unnatural substrate 3-O-methylfluorescein phosphate (OMFP). However, we observed residual in vitro activity of D72A against PI(4,5)P(2), and in accordance with this, it exhibits the same cellular phenotype as WT PRL-3. Our analysis of the A111S variant shows that the hyperactivity toward the unnatural OMFP substrate is not apparent in dephosphorylation assays with phosphoinositides: the mutant is completely inactive against PIPs. We observed significant structural destabilization of this variant. The cellular phenotype of this mutant equals that of the catalytically inactive C104S mutant. These results provide a possible explanation for the absence of the conserved Ser of the PTP catalytic motif in the PRL family. The correlation of the phosphatase activity toward PI(4,5)P(2) with the observed phenotypes for WT PRL-3 and the mutants suggests a link between the PI(4,5)P(2) dephosphorylation by PRL-3 and its role in cell migration.