Crystal structures of two nucleic acid-binding proteins

S1 nuclease from Aspergillus oryzae is a glycoprotein of 32 kDa molecular weight. The protein has two enzymatic activities: it is an endo-exonuclease with high specificity for single stranded nucleic acids, and it has an additional 3' -nucleotidase activity. S1 nuclease is widely used in molecular biology as a single-strand specific nuclease due to its high stability and efficiency. It cleaves single-stranded regions of nucleic acids producing 5' -nucleotides without significant side-reactions. The crystal structure of S1 nuclease has been determined to 1.7 Å resolution by molecular replacement based on the known structure of PI nuclease from Penicillinum citrinum, which has 49 % sequence identity compared to S1. The overall fold and the active site of S1 nuclease is basically identical to that of PI nuclease, and also very similar to Phospholipase C from Bacillus cereus and alpha-toxin from Clostridium perfringens. The characteristic feature of this family of enzymes is a trinuclear zinc cluster in their active sites. A BLAST search in the sequence databases revealed several other protein sequences from bacteria, protozoa and plants possessing an approximately 30% sequence identity compared to S1 nuclease, but showing an almost complete conservation of structurally and functionally important residues. Soaking and co-crystallisation experiments with substrate analogues have been carried out in order to obtain an enzyme-substrate complex. These efforts have not resulted in the structure determination of any complexes under crystallisation conditions: no binding of substrate has been observed. Nevertheless, an enzyme mechanism has been proposed based on structural data of S1 nuclease and nucleases with similar active sites. In eukaryotes Sm and Sm-like proteins are the core components of the small nuclear ribonucleoprotein particles (snRNPs), which are involved in a variety of functions including rRNA processing, tRNA maturation and pre-mRNA processing. The Sm proteins are 70 to 120 amino acids long and share a common bi-partite signature sequence. The spliceosome, where the transesterification reaction of splicing occurs, is assembled by several snRNPs named after their constituting snRNA: U1, U2, U4, U5 and U6. An snRNA contains a short single stranded, uridine rich sequence motif, where the Sm proteins bind, but the three-dimensional arrangement of the Sm proteins and the mode of binding is unknown. In humans there are seven different canonical Sm proteins, which according to biochemical and electron microscopic studies seem to form a seven membered ring in vitro. Recently two crystal structures of human Sm protein dimers have been published. Interestingly Sm-related protein sequences have been found in the available genomic database of various Archaebacteria based on sequence homology. In contrast with eukaryotes only one or two Sm-related protein sequences have been identified in one organism. Their function is currently unknown, since analogous pre-mRNA splicing does not occur in Archaebacteria. Two Sm-related proteins of Archaeoglobus fulgidus have been cloned and expressed as fusion proteins. One of them called AF-Sm2 has been o crystallised utilising ammonium sulphate as precipitant and solved to 1.95 Å resolution by SIRAS using a single mercury derivative. AF-Sm2 crystallises in a hexagonal space group (P6) and contains one molecule per asymmetric unit. The 77 residue long protein has a very similar fold compared to the solved human Sm protein structures: a short N-terminal α-helix followed by a five stranded, strongly bent, U-shaped β-sheet resulting in a barrel-like overall fold. Six AF-Sm2 molecules form a ring in the crystal structure mediated by extensive hydrophobic and hydrogen-bonding interactions. Gel filtration experiments have indicated a pH dependence of oligomerisation in accordance with the crystallisation experiences. Currently the target of the Sm-related proteins of Archaeoglobus fulgidus and the stochiometry of oligomerisation in vivo is completely unknown