Search

Trans-activator (tat) proteins are necessary components for the completion of the T replication cycle of lentiviruses. The three-dimensional structure of the equine infectious anemia virus (EIAV) tat protein (e-tat) was studied with CD spectroscopy, NMR spectroscopy, and restrained molecular-dynamics calculations. No stable elements of regular secondary structure were detected, but the sequence regions responsible for nucleic acid binding showed helix-forming tendency, e-tat exhibits a flexible tertiary structure, and only the amino acids comprising the core sequence region form a well-defined tertiary fold. The three-dimensional structure allows discussion of biochemical data as well as data from molecular biological investigations of lentiviral tat proteins.

The solution structure of the 60-residue 1[Fe4-S4] ferredoxin from the hyperthermophilic bacterium Thermotoga maritima was determined based on 683 distance and 35 dihedral angle restraints that were obtained from NMR data. In addition, data known from crystallographic studies of ferredoxins was used for modeling of the iron-sulfur cluster and its environment. The protein shows a globular fold very similar to the fold of the related 1[Fe4-S4] ferredoxins from Desulfovibrio gigas and Desulfovibrio africanus, and elements of regular secondary structure similar to those in other ferredoxins were found in the T. maritima protein. In particular, the T. maritima protein displayed a beta-sheet structure made up of strands located at the very NH(2) and COOH termini of the protein, and an internal alpha-helix. The internal beta-sheet observed in the D. gigas and D. africanus ferredoxins could not be confirmed in T. maritima ferredoxin and is thus suggested to be only weakly present or even absent in this protein. This result suggests that thermostability in ferredoxins is not necessarily correlated with the content of stable elements of regular secondary structure.

A 72-amino-acid chimeric protein, Chi1, was constructed from the N-terminal part of interleukin 8, IL-8-(1-53), and the C-terminal part of melanoma growth stimulatory activity, MGSA-(54-72). Chi1 protein showed receptor-binding specificity and biological activity similar, but not identical to IL-8 and decidedly different from MGSA. The structure of Chi1 was determined in solution by two-dimensional NMR and molecular-dynamics calculations. The structure resembled the structures of MGSA and IL-8 closely, containing a triple-stranded beta-sheet in the IL-8 part and an amphipathic alpha-helix in the MGSA part. Chi1 formed dimers at millimolar concentrations via the first strand from the N-terminus, analogous to IL-8 and MGSA. In contrast to the latter molecules, however, the alpha-helix of Chi1 did not pack against the beta-sheet part, but was an independent structural element. This structural difference could be explained mainly by the modulation of hydrophobic interactions between the helix and the rest of the protein in Chi1 as compared to IL-8 and MGSA. It is concluded that tight helix packing is not required for receptor binding and biological activity of Chi1.

One of the principle peptide components of the amyloid plaque deposits of Alzheimer's disease in humans is the 40-amino-acid peptide beta-amyloid A4-(1-40)-peptide. The full-length A4-(1-40)-peptide was chemically synthesized and the solution structure determined by two-dimensional nuclear magnetic resonance spectroscopy and restrained molecular-dynamics calculations. Synthetic human A4-(1-40)-peptide was soluble and non-aggregating for several days in 40% (by vol.) trifluoroethanol/water. All spin systems could be unambiguously assigned, and a total of 203 sequential and medium-range cross-peaks were found in the NOESY (nuclear Overhauser enhancement spectroscopy) spectrum. Long-range NOE cross-peaks that would indicate tertiary structure of the peptide were absent. The main secondary-structure elements found by chemical-shift analysis, sequential and medium-range NOESY data, and NOE-based restrained molecular-dynamics calculations were two helices, Gln15-Asp23 and Ile31-Met35, whereas the rest of the peptide was in random-coil conformation. A similar secondary structure is suggested for the aggregation part of prions, the postulated causative agents of the transmissible spongiform encephalopathy. The sequence of the helical part of prion proteins was observed to be remarkably similar to the sequence of the helical part of human A4-(1-40)-peptide.

The solution structure of a 15-mer nutRboxB RNA hairpin complexed with the 36-mer N-terminal peptide of the N protein (N36) from bacteriophage lambda was determined by 2D and 3D homonuclear and heteronuclear magnetic resonance spectroscopy. These 36 amino acids include the arginine-rich motif of the N protein involved in transcriptional antitermination of phage lambda. Upon complex formation with boxB RNA, the synthetic N36 peptide binds tightly to the major groove of the boxB hairpin through hydrophobic and electrostatic interactions forming a bent alpha helix. Four nucleotides of the GAAAA pentaloop of the boxB RNA adopt a GNRA-like tetraloop fold in the complex. The formation of a GAAA tetraloop involves a loop-closing sheared base pair (G6-A10), base stacking of three adenines (A7, A8, and A10), and extrusion of one nucleotide (A9) from the loop, as observed previously for the complex of N(1-22) peptide and the nutLboxB RNA [Legault, P., Li, J., Mogridge, J., Kay, L.E.Greenblatt, J. (1998) Cell 93, 289-299]. Stacking of the bases is extended by the indole-ring of Trp18 which also forms hydrophobic contacts to the side-chains of Leu24, Leu25, and Val26. Based on the structure of the complex, three mutant peptides were synthesized and investigated by CD and NMR spectroscopy in order to determine the role of particular residues for complex formation. These studies revealed very distinct amino-acid requirements at positions 3, 4, and 8, while replacement of Trp18 with tyrosine did not result in any gross structural changes.

The solution structure of the 75-amino-acid trans-activator (Tat) protein of the equine infectious-anemia virus in trifluoroethanol-containing solution was determined by two-dimensional and three-dimensional nuclear magnetic resonance spectroscopy, resulting in a total of 838 nuclear-Over-hauser-enhancement distance restraints, and restrained molecular-dynamics simulations. In contrast to the recently determined structure of this protein in trifluoroethanol-free pH 6.3 solution, the hydrophobic core and the adjacent basic RNA-binding region of the protein showed well-defined alpha-helical secondary structure in trifluoroethanol-containing solution. The helical regions comprise those parts of the molecule whose helix-forming tendencies were noted earlier in trifluoroethanol-free solution. Two helices (Gln38-Arg43 and Asp48-Ala64) are connected by a tight type-II turn centered at the strictly conserved Gly46 leading to a helix-turn-helix motif in the core and basic region of the protein. A third helix (Thr9-Asn13) is located in the less well defined N-terminal part of the protein. These observations may support the notion that the protein adopts a helical structure in the RNA-binding region on complex formation. Although the secondary-structure elements become better defined in trifluoroethanol-containing solution, the opposite is true for the hydrophobically stabilized tertiary structure. This adds a caveat to studies of protein structures in trifluoroethanol-containing solution in general.

Nuclear magnetic resonance (NMR) spectroscopy revealed features of the secondary structure of the equine infectious anemia virus (EIAV) Tat protein in solution. We could show that this protein, which is required in the replication cycle of lentiviruses, forms a predominantly helical structure in trifluoroethanol/water (40% by vol.) solution. In particular, the basic RNA-binding region and the adjacent core domain, which are highly conserved among lentiviral Tat proteins, show helix-type secondary structure under these conditions. Our observations, in concert with recent biochemical data from other laboratories, suggest that the core sequence region and the basic sequence region form interdependent structural domains, both possibly necessary for correct RNA binding.