M.TaqI: possible catalysis via cation-pi interactions in N-specific DNA methyltransferases.

The adenine-specific DNA methyltransferase M.TaqI transfers a methyl group from S-adenosylmethionine to N6 of the adenine residue in the DNA sequence 5'-TCGA-3'. In the crystal structure of M.TaqI in complex with S-adenosylmethionine the enzyme is folded into two domains: An N-terminal catalytic domain, whose fold is conserved among S-adenosyl-methionine dependent methyltransferases, and a DNA recognition domain which possesses a unique fold. In the active site, two aromatic residues, Tyr 108 and Phe 196, are postulated to bind the flipped-out target DNA adenine which becomes methylated. By lowering the energy of the positively charged transition state via cationic-pi interactions, these two residues probably hold a key role in catalysis.

X-ray crystal structure of arrestin from bovine rod outer segments.

Retinal arrestin is the essential protein for the termination of the light response in vertebrate rod outer segments. It plays an important role in quenching the light-induced enzyme cascade by its ability to bind to phosphorylated light-activated rhodopsin (P-Rh*). Arrestins are found in various G-protein-coupled amplification cascades. Here we report on the three-dimensional structure of bovine arrestin (relative molecular mass, 45,300) at 3.3 A resolution. The crystal structure comprises two domains of antiparallel beta-sheets connected through a hinge region and one short alpha-helix on the back of the amino-terminal fold. The binding region for phosphorylated light-activated rhodopsin is located at the N-terminal domain, as indicated by the docking of the photoreceptor to the three-dimensional structure of arrestin. This agrees with the interpretation of binding studies on partially digested and mutated arrestin.

Crystallization and preliminary X-ray analysis of arrestin from bovine rod outer segment.

We present the first X-ray study of a member of the arrestin family, the bovine retinal arrestin. Arrestin is essential for the fine regulation and termination of the light-induced enzyme cascade in vertebrate rod outer segments. It plays an important role in quenching phototransduction by its ability to preferentially bind to phosphorylated light-activated rhodopsin. The crystals diffract between 3 angstroms and 3.5 angstroms (space group P2(1)2(1)2, cell dimensions a = 169.17 angstroms, b = 185.53 angstroms, c = 90.93 angstroms, T = 100 K). The asymmetric unit contains four molecules with a solvent content of 68.5% by volume.

A model for DNA binding and enzyme action derived from crystallographic studies of the TaqI N6-adenine-methyltransferase.

The crystal structures of the DNA-N6-adenine-methyltransferase M.TaqI, in complexes with the cofactor S-adenosyl-L-methionine (AdoMet) and the competitive inhibitor sinefungin (Sf) show identical folding of the polypeptide chains into two domains. The N-terminal domain carries the cofactor-binding site, the C-terminal domain is thought to be implicated in sequence-specific DNA binding. Model building of the M.TaqI-DNA complex suggests that the adenine to be methylated swings out of the double helix as found previously in the cytosine-C5-MTase HhaI DNA co-crystal structure. A torsion of the methionine moiety of the cofactor is required to bring the methyl group within reach of the swung-out base and allow methyl group transfer.

Three-dimensional structure of the adenine-specific DNA methyltransferase M.Taq I in complex with the cofactor S-adenosylmethionine.

The Thermus aquaticus DNA methyltransferase M.Taq I (EC 2.1.1.72) methylates N6 of adenine in the specific double-helical DNA sequence TCGA by transfer of --CH3 from the cofactor S-adenosyl-L-methionine. The x-ray crystal structure at 2.4-A resolution of this enzyme in complex with S-adenosylmethionine shows alpha/beta folding of the polypeptide into two domains of about equal size. They are arranged in the form of a C with a wide cleft suitable to accommodate the DNA substrate. The N-terminal domain is dominated by a nine-stranded beta-sheet; it contains the two conserved segments typical for N-methyltransferases which form a pocket for cofactor binding. The C-terminal domain is formed by four small beta-sheets and alpha-helices. The three-dimensional folding of M.Taq I is similar to that of the cytosine-specific Hha I methyltransferase, where the large beta-sheet in the N-terminal domain contains all conserved segments and the enzymatically functional parts, and the smaller C-terminal domain is less structured.

X-ray crystallographic and calorimetric studies of the effects of the mutation Trp59-->Tyr in ribonuclease T1.

Two mutants of ribonuclease T1 (RNaseT1), [59-tyrosine]ribonuclease T1 (W59Y) and [45-tryptophan,59-tyrosine]ribonuclease T1 (Y45W/W59Y) possess between 150% and 190% wild-type activity. They have been crystallised as complexes of the inhibitor 2'-guanylic acid and analysed by X-ray diffraction at resolutions of 0.23 nm and 0.24 nm, respectively. The space group for both is monoclinic, P2(1), with two molecules/asymmetric unit, W59Y: a = 4.934 nm, b = 4.820 nm, c = 4.025 nm, beta = 90.29 degrees. Y45W/W59Y: a = 4.915 nm, b = 4.815 nm, c = 4.015 nm, beta = 90.35 degrees. Compared to wild-type RNaseT1 in complex with 2'-guanylic acid (2'GMP) both mutant inhibitor complexes indicate that the replacement of Trp59 by Tyr leads to a 0.04-nm inward shift of the single alpha-helix and to significant differences in the active-site geometry, inhibitor conformation and inhibitor binding. Calorimetric studies of a range of mutants [24-tryptophan]ribonuclease T1 (Y24W), [42-tryptophan]ribonuclease T1 (Y42W), [45-tryptophan]ribonuclease T1 (Y45W), [92-alanine]ribonuclease T1 (H92A) and [92-threonine]ribonuclease T1 (H92T) with and without the further mutation Trp59-->Tyr showed that mutant proteins for which Trp59 is replaced by Tyr exhibit slightly decreased thermal stability.

Three-dimensional structure of the ternary complex between ribonuclease T1, guanosine 3',5'-bisphosphate and inorganic phosphate at 0.19 nm resolution.

The ternary complex formed between RNase T1, guanosine 3',5'-bisphosphate (3',5'-pGp) and Pi crystallizes in the cubic space group I23 with a = 8.706(1) nm. In a previous publication [Lenz, A., Heinemann, U., Maslowska, M. & Saenger, W. (1991) Acta Crystallogr. B47, 521-527], the structure of the complex (in which Pi was not located) was described at a resolution of 0.32 nm. This is now extended to 0.19 nm with newly grown, larger crystals. Refinement with restrained least-squares converged at R = 17.8% for 8027 reflections with [Fo[ > or = 1 sigma ([Fo[); the final model comprises 120 water molecules. 3',5'-pGp is bound to RNase T1 in the anti form, with guanine in the specific recognition site; the 3'-phosphate protrudes into the solvent, and the 5'-phosphate hydrogen bonds with Lys41 O and Asn43 N4. A tetrahedral anion assigned as Pi occupies the catalytic site and hydrogen bonds to the side chains of Tyr38, Glu58, Arg77 and His92. The overall polypeptide fold of RNase T1 in the cubic space group does not differ significantly from that in the orthorhombic space group P2(1)2(1)2(1) except for changes < or = 0.2 nm in loop regions 69-72 and 95-98.

Crystal structure of the factor for inversion stimulation FIS at 2.0 A resolution.

The factor for inversion stimulation (FIS) binds as a homodimeric molecule to a loose 15 nucleotide consensus sequence in DNA. It stimulates DNA-related processes, such as DNA inversion and excision, it activates transcription of tRNA and rRNA genes and it regulates its own synthesis. FIS crystallizes as a homodimer, with 2 x 98 amino acid residues in the asymmetric unit. The crystal structure was determined with multiple isomorphous replacement and refined to an R-factor of 19.2% against all the 12,719 X-ray data (no sigma-cutoff) extending to 2.0 A resolution. The two monomers are related by a non-crystallographic dyad axis. The structure of the dimer is modular, with the first 23 amino acid residues in molecule M1 and the first 24 in molecule M2 disordered and not "seen" in the electron density. The polypeptide folds into four alpha-helices, with alpha A, alpha A' (amino acid residues 26 to 40) and alpha B, alpha B' (49 to 69) forming the core of the FIS dimer, which is stabilized by hydrophobic forces. To the core are attached "classical" helix-turn-helix motifs, alpha C, alpha D (73 to 81 and 84 to 94) and alpha C', alpha D'. The connections linking the helices are structured by two beta-turns for alpha A/alpha B, and alpha C1 type extensions are observed at the C termini of helices alpha B, alpha C and alpha D. Helices alpha D and alpha D' contain 2 x 6 positive charges; they are separated by 24 A and can bind adjacent major grooves in B-type DNA if it is bent 90 degrees. The modular structure of FIS is also reflected by mutation experiments; mutations in the N-terminal part and alpha A interfere with FIS binding to invertases, and mutations in the helix-turn-helix motif interfere with DNA binding.

RNase T1 mutant Glu46Gln binds the inhibitors 2'GMP and 2'AMP at the 3' subsite.

On the basis of molecular dynamics and free-energy perturbation approaches, the Glu46Gln (E46Q) mutation in the guanine-specific ribonuclease T1 (RNase T1) was predicted to render the enzyme specific for adenine. The E46Q mutant was genetically engineered and characterized biochemically and crystallographically by investigating the structures of its two complexes with 2'AMP and 2'GMP. The ribonuclease E46Q mutant is nearly inactive towards dinucleoside phosphate substrates but shows 17% residual activity towards RNA. It binds 2'AMP and 2'GMP equally well with dissociation constants of 49 microM and 37 microM, in contrast to the wild-type enzyme, which strongly discriminates between these two nucleotides, yielding dissociation constants of 36 microM and 0.6 microM. These data suggest that the E46Q mutant binds the nucleotides not to the specific recognition site but to the subsite at His92. This was confirmed by the crystal structures, which also showed that the Gln46 amide is hydrogen bonded to the Phe100 N and O atoms, and tightly anchored in this position. This interaction may either have locked the guanine recognition site so that 2'AMP and 2'GMP are unable to insert, or the contribution to guanine recognition of Glu46 is so important that the E46Q mutant is unable to function in recognition of either guanine and adenine.

Structure of ribonuclease T1 complexed with zinc(II) at 1.8 A resolution: a Zn2+.6H2O.carboxylate clathrate.

In order to study the inhibitory effect of Zn2+ on ribonuclease T1 [RNase T1; Itaya & Inoue (1982). Biochem. J. 207, 357-362], the enzyme was cocrystallized with 2 mM Zn2+, pH 5.2, from a solution containing 55% (v/v) 2-methyl-2,4-pentanediol. The crystals are orthorhombic, P2(1)2(1)2(1), a = 48.71 (1), b = 46.51 (1), c = 41.14 (1) A, Z = 4, V = 93203 A3. The crystal structure was determined by molecular replacement and refined by restrained least-squares methods based on Fhkl for 8291 unique reflections with Fo greater than or equal to 1 sigma (Fo) in the resolution range 10 to 1.8 A and converged at a crystallographic R factor of 0.140. The Zn2+ is not bonded to the active site of RNase T1, probably because the His40 and His92 side chains are protonated. Zn2+ occupies the same site as Ca2+ in a series of crystal structures of free and nucleotide-complexed RNase T1. It is coordinated to Asp15 carboxylate and to six water molecules forming a dodecahedron of square antiprismatic form. The Zn2+...O distances are approximately 2.5 A, suggesting that Zn2+ is clathrated and not coordinated, which would require distances of 2.0 A.

Crystallization of the hybrid Bacillus (1-3,1-4)-beta-glucanase H(A16-M).

The extremely thermostable hybrid Bacillus (1-3,1-4)-beta-glucanase H(A16-M) has been crystallized with polyethylene glycol by vapour diffusion. The single crystals diffract to a resolution of 2.2 A. The protein crystallizes in the orthorhombic space group P2(1)2(1)2(1) with a = 70.22(7) A, b = 72.56(1) A, c = 49.97(1) A, and has one molecule per asymmetric unit.

Three-dimensional structure of the E. coli DNA-binding protein FIS.

The factor for inversion stimulation, FIS, is involved in several cellular processes, including site-specific recombination and transcriptional activation. In the reactions catalysed by the DNA invertases Gin, Hin and Cin, FIS stimulates recombination by binding to an enhancer sequence. Within the enhancer, two FIS dimers (each 2 x 98 amino acids) bind to two 15-base-pair consensus sequences and induce bending of the DNA. Current models propose that the enhancer-FIS complex organizes a specific synapse, either through direct interactions with Gin, or by modelling the substrate into a configuration suitable for recombination. Using X-ray analysis at 2.0 A resolution, we now show that FIS is composed of four alpha helices tightly intertwined to form a globular dimer with two protruding helix-turn-helix motifs. The 24 N-terminal amino acids are so poorly defined in the electron density map as to make interpretation doubtful, indicating that they might act as 'feelers' suitable for DNA or protein (invertase) recognition. We infer from model building that DNA has to bend for tight binding to FIS.