A quantitative survey on the performance of multireference (MR), con®guration interaction with all singles and doubles (CISD), MRCISD with the Davidson correction and MR-average quadratic coupled cluster (AQCC) methods for a wide range of excited states of the diatomic molecules B 2 , C 2 , N 2 and O 2 is presented. The spectroscopic constants r e , x e , T e and D e for a total of 60 states have been evaluated and critically compared with available experimental data. Basis set extrapolations and size-extensivity corrections are essential for highly accurate results: MR-AQCC mean-errors of 0.001 A Ê , 10 cm À1 , 300 cm À1 and 300 cm À1 have been obtained for r e , x e , T e and D e , respectively. Owing to the very systematic behavior of the results depending on the basis set and the choice of method, shortcomings of the calculations, such as Rydberg state coupling or insucient con®guration spaces, can be identi®ed independently of experimental data. On the other hand, signi®cant discrepancies with experiment for states which indicate no shortcomings whatsoever in the theoretical treatment suggest the re-evaluation of experimental results. The broad variety of states included in our survey and the uniform quality of the results indicate that the observed systematics is a general feature of the methods and, hence, is molecule-independent.

dUTPase is responsible for preventive DNA repair via exclusion of uracil. Developmental regulation of the Drosophila enzyme is suggested to be involved in thymine-less apoptosis. Here we show that in addition to conserved dUTPase sequence motifs, the gene of Drosophila enzyme codes for a unique Ala-Pro-rich segment. Kinetic and structural analyses of the recombinant protein and a truncation mutant show that the Ala-Pro segment is flexible and has no regulatory role in vitro. The homotrimer enzyme unfolds reversibly as a trimeric entity with a melting temperature of 54°C, 23°C lower than Escherichia coli dUTPase. In contrast to the bacterial enzyme, Mg 2؉ binding modulates conformation of fly dUTPase, as identified by spectroscopy and by increment in melting temperature. A single well folded, but inactive, homotrimeric core domain is generated through three distinct steps of limited trypsinolysis. In fly, but not in bacterial dUTPase, binding of the product dUMP induces protection against proteolysis at the tryptic site reflecting formation of the catalytically competent closed conformer. Crystallographic analysis argues for the presence of a stable monomer of Drosophila dUTPase in crystal phase. The significant differences between prototypes of eukaryotic and prokaryotic dUTPases with respect to conformational flexibility of the active site, substrate specificity, metal ion binding, and oligomerization in the crystal phase are consistent with alteration of the catalytic mechanism and hydropathy of subunit interfaces.

The catalytic mechanism of dUTP pyrophosphatase (dUTPase), responsible for the prevention of uracil incorporation into DNA, involves ordering of the flexible C terminus of the enzyme. This conformational shift is investigated by multidimensional NMR on the Drosophila enzyme. Flexible segments of the homotrimer give rise to sharp resonances in the (1)H-(15)N heteronuclear single-quantum coherence (HSQC) spectra, which are clearly distinguishable from the background resonances of the well folded protein globule. Binding of the product dUMP or the analogues dUDP and alpha,beta-imino-dUTP to the enzyme induces a conformational change reflected in the disappearance of eight sharp resonances. This phenomenon is interpreted as nucleotide binding-induced ordering of some residues upon the folded protein globule. Three-dimensional (15)N-edited (1)H-(15)N HSQC total correlation spectroscopy (TOCSY) and (1)H-(15)N HSQC nuclear Overhauser effect spectroscopy measurements allowed clear assignment of these eight specific resonance peaks. The residues identified correspond to the conserved C-terminal sequence motif, indicating that (i) this conformational shift is amenable to NMR studies in solution even in the large trimeric molecule and (ii) formation of the closed enzyme conformer in the case of the Drosophila enzyme does not require the complete triphosphate chain of the substrate. NMR titration of the enzyme with the nucleotide ligands as well as kinetic data indicated significant deviation from the model of independent active sites within the homotrimer. The results suggest allosterism in the eukaryotic dUTPase.