Two highly similar activities are represented in this group: thymidine phosphorylase (TP, gene deoA, EC: 2.4.2.4) and pyrimidine-nucleoside phosphorylase (PyNP, gene pdp, EC: 2.4.2.2). Both are dimeric enzymes that function in the salvage pathway to catalyse the reversible phosphorolysis of pyrimidine nucleosides to the free base and sugar moieties. In the case of thymidine phosphorylase, thymidine (and to a lesser extent, 2'-deoxyuridine) is lysed to produce thymine (or uracil) and 2'-deoxyribose-1-phosphate. Pyrimidine-nucleoside phosphorylase performs the analogous reaction on thymidine (to produce the same products) and uridine (to produce uracil and ribose-1-phosphate). PyNP is typically the only pyrimidine nucleoside phosphorylase encoded by Gram positive bacteria, while eukaryotes and proteobacteria encode two: TP, and the unrelated uridine phosphorylase. In humans, TP was originally characterized as platelet-derived endothelial cell growth factor and gliostatin [PMID:9310231]. Structurally, the enzymes are homodimers, each composed of a rigid all alpha-helix lobe and a mixed alpha-helix/beta-sheet lobe, which are connected by a flexible hinge [PMID:9698549, PMID:9817849]. Prior to substrate binding, the lobes are separated by a large cleft. A functional active site and subsequent catalysis occurs upon closing of the cleft. The active site, composed of a phosphate binding site and a (deoxy)ribonucleotide binding site within the cleft region, is highly conserved between the two enzymes of this group. Active site residues (Escherichia coli DeoA numbering) include the phosphate binding Lys84 and Ser86 (close to a glycine-rich loop), Ser113, and Thr123, and the pyrimidine nucleoside-binding Arg171, Ser186, and Lys190. Sequence comparison between the active site residues for both enzymes reveals only one difference [PMID:9817849], which has been proposed to partially mediate substrate specificity. In TP, position 111 is a methionine, while the analogous position in PyNP is lysine. It should be noted that the uncharacterised archaeal members of this family differ in a number of respects from either of the characterized activities. The residue at position 108 is lysine, indicating the activity might be PyNP-like (though the determinants of substrate specificity have not been fully elucidated). Position 171 is glutamate (negative charge side chain) rather than arginine (positive charge side chain). In addition, a large loop that may "lock in" the substrates within the active site is much smaller than in the characterized members. It is not clear what effect these and other differences have on activity and specificity.