Technical Highlight - October 2009
Short description: Many annotations of the cytosolic sulfotransferase in publicly available databases are found to be wrong.
In early 2001, a small team from the National Institutes of Health published a paper 1 providing biochemical evidence that all cytosolic sulfotransferases that had been published to that date had the same protein–protein interaction motif. Petrotchenko et al. found through cross-linking, protease digestion and mass spectrometry that these proteins, which catalyze the sulfonation of hormones, metabolites and other substrates, form homo- and heterodimers through an unusually small dimer interface observed in their protein crystals
This important biological information seems to have been subsequently ignored or forgotten by the majority of crystallographers working on these sulfotransferases, despite biological evidence to support this interface. Since the paper was published 1 , a further 21 structures have been deposited in the PDB, of which only three have an annotation about this interface probably being the physiologically relevant interaction site.
Was this because the interface is absent in the other structures? Brian Weitzner et al. 2 from PSI NMHRCM and Fox Chase Cancer Center, Philadelphia, analysed all 28 structures and found that wasn't the case. All the structures, in 17 crystal forms, contain the same dimeric structure. Surprisingly, even the crystal form of mouse SULTIE1, which is monomeric in solution experiments 1 , contains the common dimeric interface in its crystal, although it is distorted and missing two important salt bridges.
When Weitzner et al. checked the annotations in the publicly available databases such as the Protein Databank (PDB), the Protein Quaternary Server (PQS), and Protein Interfaces, Surfaces and Assemblies (PISA), they found that for almost all of the structures, the servers did not indicate that the small Petrochenko interface was the likely dimer and sometimes instead provided dimer structures unique to each crystal, which are unlikely to be biologically relevant. Structures of mouse SULT1D1 published after the Weitzner paper also do not have annotations showing the Petrotchenko dimer, even though it is present in the crystals (PDB entries 2ZPT, 2ZVP and others).
Although crystallographers routinely check the molecular weight of a protein in roughly physiological conditions to assess its oligomerization state, not so many routinely test the interfaces. The Dunbrack team is now working on software that can analyze the whole of the PDB to help researchers identify the biologically relevant interaction site.
E. V. Petrotchenko, L. C. Pedersen, C. H. Borchers, K. B. Tomer & M. Negishi The dimerization motif of cytosolic sulfotransferases.
FEBS Lett. 490, 39-43 (2001). doi:10.1016/S0014-5793(01)02129-9
B. Weitzner, T. Meehan, Q. Xu & R. L. Dunbrack, Jr An unusually small dimer interface is observed in all available crystal structures of cytosolic sulfotransferases.
Proteins 75, 289-295 (2009). doi:10.1002/prot.22347