Important structural requirements of the catalytic site of protein serine/threonine phosphatases are to bind serine-phosphate and threonine-phosphate. The results of a survey of the structures of 30 phosphate binding sites in 18 different proteins were reported by Johnson in 1984 [45]. It was shown that phosphate binding sites can be separated into those involved solely in binding and those present at the catalytic site. The latter sites are generally less well defined. Arginine is frequently involved in binding phosphate with lysine less often. The phosphate binding sites are commonly present at the N-terminus of an alpha-helix to benefit from positive helix dipole interactions. Phosphate groups also interact with the side chains of serine, threonine, asparagine and glutamine and with the NH peptide groups of the main-chain. Glycines residues are common at these positions.
Amongst the eukaryotic protein serine/threonine phosphatase sequences (Figure 1), Arg is invariant at positions 92, 136, 207, 268 and 293. However, when the bacteriophage sequences are included only positions 92 and 136 are invariant, and on including the diadenosine tetra-phosphatase sequence only position 92 is invariant. This strongly suggests a role for Arg 92 in phosphate binding and less so for Arg 136. Arg 92 and 136 occur in highly conserved regions of the sequence predicted as loops. Asn 138, 324, 331; Gln 63; Ser 96 and 321 and Lys 125 are invariant amongst the eukaryotic sequences, although only asparagine 138 is conserved throughout all the phosphatase sequences. Asn 138 is therefore a candidate for a phosphate binding residue. There are 11 invariant glycine residues in the eukaryotic sequences, many of which are located close in primary structure to invariant arginine, aspartate and histidine residues, suggesting a role in determining the position of these residues on surface turns and loops. Amongst eukaryotic and bacteriophage phosphatases and the diadenosine tetra-phosphatase sequences, glycine is invariant at positions 58, 62, 87, 93 and 137 and it is possible that one or more of these glycines is involved in phosphate binding.
The secondary structure prediction (Figure 1) suggests that the region
between 54 and 98 has the pattern -
-. This is similar but
not identical, to the phosphate-binding structure of the dehydrogenase
protein family. In the dehydrogenases, the phosphatate binds near the
N-terminus of the first
- helix in a
-
-
- motif where all
- strands are parallel, and phosphate binding is
stabilised by the
-helix dipole [46]. In the phosphatases,
if the region 54-98 were to fold as a parallel
-
- unit with an
additional antiparallel strand at the end of the sheet, then the
-
-loop (58-63) that contains the invariant residue pattern
GD[IVLTL]HG, and the
-
-loop which contains the pattern
GD[LYF]V[DA]RG would be suitably positioned to form a phosphate
binding site with stabilisation from the dipole of the intervening
helix (64-73).