Three-dimensional structure of a hammerhead ribozyme.

The hammerhead ribozyme is a small catalytic RNA motif made up of three base-paired stems and a core of highly conserved, non-complementary nucleotides essential for catalysis. The X-ray crystallographic structure of a hammerhead RNA-DNA ribozyme-inhibitor complex at 2.6 A resolution reveals that the base-paired stems are A-form helices and that the core has two structural domains. The first domain is formed by the sequence 5'-CUGA following stem I and is a sharp turn identical to the uridine turn of transfer RNA, whereas the second is a non-Watson-Crick three-base-pair duplex with a divalent-ion binding site. The phosphodiester backbone of the DNA inhibitor strand is splayed out at the phosphate 5' to the cleavage site. The structure indicates that the ribozyme may destabilize a substrate strand in order to facilitate twisting of the substrate to allow cleavage of the scissile bond.

Model for an RNA tertiary interaction from the structure of an intermolecular complex between a GAAA tetraloop and an RNA helix.

In large structured RNAs, RNA hairpins in which the strands of the duplex stem are connected by a tetraloop of the consensus sequence 5'-GNRA (where N is any nucleotide, and R is either G or A) are unusually frequent. In group I introns there is a covariation in sequence between nucleotides in the third and fourth positions of the loop with specific distant base pairs in putative RNA duplex stems: GNAA loops correlate with successive 5'-C-C.G-C base pairs in stems, whereas GNGA loops correlate with 5'-C-U.G-A. This has led to the suggestion that GNRA tetraloops may be involved in specific long-range tertiary interactions, with each A in position 3 or 4 of the loop interacting with a C-G base pair in the duplex, and G in position 3 interacting with a U-A base pair. This idea is supported experimentally for the GAAA loop of the P5b extension of the group I intron of Tetrahymena thermophila and the L9 GUGA terminal loop of the td intron of bacteriophage T4 (ref. 4). NMR has revealed the overall structure of the tetraloop for 12-nucleotide hairpins with GCAA and GAAA loops and models have been proposed for the interaction of GNRA tetraloops with base pairs in the minor groove of A-form RNA. Here we describe the crystal structure of an intermolecular complex between a GAAA tetraloop and an RNA helix. The interactions we observe correlate with the specificity of GNRA tetraloops inferred from phylogenetic studies, suggesting that this complex is a legitimate model for intramolecular tertiary interactions mediated by GNRA tetraloops in large structured RNAs.

Crystals of a hammerhead ribozyme.

Hammerhead ribozyme-inhibitor complexes consisting of an RNA "enzyme" strand and a DNA "inhibitor" strand have been crystallized in four different crystal forms. The crystal form that is most attractive for structure determination diffracts to approximately 3.2 A resolution; it is trigonal, space group P3(1)21 or enantiomorph, a = 92.5 A, c = 185.0 A.

Structural comparison suggests that thermolysin and related neutral proteases undergo hinge-bending motion during catalysis.

Crystal structures are known for three members of the bacterial neutral protease family: thermolysin from Bacillus thermoproteolyticus (TLN), the neutral protease from Bacillus cereus (NEU), and the elastase of Pseudomonas aeruginosa (PAE), both in free and ligand-bound forms. Each enzyme consists of an N-terminal and C-terminal domain with the active site formed at the junction of the two domains. Comparison of the different molecules reveals that the structure within each domain is well conserved, but there are substantial hinge-bending displacements (up to 16 degrees) of one domain relative to the other. These domain motions can be correlated with the presence or absence of bound inhibitor, as was previously observed in the specific example of PAE [Thayer, M.M., Flaherty, K.M., & McKay, D.B. (1991) J. Biol. Chem. 266, 2864-2871]. The binding of inhibitor appears to be associated with a reduction of the domain hinge-bending angle by 6-14 degrees and a closure of the "jaws" of the active site cleft by about 2 A. Crystallographic refinement of the structure of thermolysin suggests that electron density seen in the active site of the enzyme in the original structure determination probably corresponds to a bound dipeptide. Thus, the crystal structure appears to correspond to an enzyme-inhibitor or enzyme-product complex, rather than the free enzyme, as has previously been assumed.