The TB structural genomics consortium: a resource for Mycobacterium tuberculosis biology

The TB Structural Genomics Consortium is an organization devoted to encouraging, coordinating, and facilitating the determination and analysis of structures of proteins from Mycobacterium tuberculosis. The Consortium members hope to work together with other M. tuberculosis researchers to identify M. tuberculosis proteins for which structural information could provide important biological information, to analyze and interpret structures of M. tuberculosis proteins, and to work collaboratively to test ideas about M. tuberculosis protein function that are suggested by structure or related to structural information. This review describes the TB Structural Genomics Consortium and some of the proteins for which the Consortium is in the progress of determining three-dimensional structures.

Solution structure of Pyrobaculum aerophilum DsrC, an archaeal homologue of the gamma subunit of dissimilatory sulfite reductase

The solution structure of DsrC, an archaeal homologue of the gamma subunit of dissimilatory sulfite reductase, has been determined by NMR spectroscopy. This 12.7-kDa protein from the hyperthermophilic archaeon Pyrobaculum aerophilum adopts a novel fold consisting of an orthogonal helical bundle with a beta hairpin along one side. A portion of the structure resembles the helix-turn-helix DNA-binding motif common in transcriptional regulator proteins. The protein contains two disulfide bonds but remains folded following reduction of the disulfides. DsrC proteins from organisms other than Pyrobaculum species do not contain these disulfide bonds. A conserved cysteine next to the C-terminus, which is not involved in the disulfide bonds, is located on a seven-residue C-terminal arm that is not part of the globular protein and is likely to dynamically sample more than one conformation.

Rapid protein-folding assay using green fluorescent protein

Formation of the chromophore of green fluorescent protein (GFP) depends on the correct folding of the protein. We constructed a "folding reporter" vector, in which a test protein is expressed as an N-terminal fusion with GFP. Using a test panel of 20 proteins, we demonstrated that the fluorescence of Escherichia coli cells expressing such GFP fusions is related to the productive folding of the upstream protein domains expressed alone. We used this fluorescent indicator of protein folding to evolve proteins that are normally prone to aggregation during expression in E. coli into closely related proteins that fold robustly and are fully soluble and functional. This approach to improving protein folding does not require functional assays for the protein of interest and provides a simple route to improving protein folding and expression by directed evolution.

Structure of translation initiation factor 5A from Pyrobaculum aerophilum at 1.75 A resolution

BACKGROUND

Translation initiation factor 5A (IF-5A) is reported to be involved in the first step of peptide bond formation in translation, to be involved in cell-cycle regulation and to be a cofactor for the Rev and Rex transactivator proteins of human immunodeficiency virus-1 and T-cell leukemia virus I, respectively. IF-5A contains an unusual amino acid, hypusine (N-epsilon-(4-aminobutyl-2-hydroxy)lysine), that is required for its function. The first step in the post-translational modification of lysine to hypusine is catalyzed by the enzyme deoxyhypusine synthase, the structure of which has been published recently.

RESULTS

IF-5A from the archebacterium Pyrobaculum aerophilum has been heterologously expressed in Escherichia coli with selenomethionine substitution. The crystal structure of IF-5A has been determined by multiwavelength anomalous diffraction and refined to 1.75 A. Unmodified P. aerophilum IF-5A is found to be a beta structure with two domains and three separate hydrophobic cores.

CONCLUSIONS

The lysine (Lys42) that is post-translationally modified by deoxyhypusine synthase is found at one end of the IF-5A molecule in an turn between beta strands beta4 and beta5; this lysine residue is freely solvent accessible. The C-terminal domain is found to be homologous to the cold-shock protein CspA of E. coli, which has a well characterized RNA-binding fold, suggesting that IF-5A is involved in RNA binding.

Formation of the ferritin iron mineral occurs in plastids

Ferritin in plants is a nuclear-encoded, multisubunit protein found in plastids; an N-terminal transit peptide targets the protein to the plastid, but the site for formation of the ferritin Fe mineral is unknown. In biology, ferritin is required to concentrate Fe to levels needed by cells (approximately 10(-7) M), far above the solubility of the free ion (10(-18) M); the protein directs the reversible phase transition of the hydrated metal ion in solution to hydrated Fe-oxo mineral. Low phosphate characterizes the solid-phase Fe mineral in the center of ferritin of the cytosolic animal ferritin, but high phosphate is the hallmark of Fe mineral in prokaryotic ferritin and plant (pea [Pisum sativum L.] seed) ferritin. Earlier studies using x-ray absorption spectroscopy showed that high concentrations of phosphate present during ferritin mineralization in vivo altered the local structure of Fe in the ferritin mineral so that it mimicked the prokaryotic type, whether the protein was from animals or bacteria. The use of x-ray absorption spectroscopy to analyze the Fe environment in pea-seed ferritin now shows that the natural ferritin mineral in plants has an Fe-P interaction at 3.26A, similar to that of bacterial ferritin; phosphate also prevented formation of the longer Fe-Fe interactions at 3.5A found in animal ferritins or in pea-seed ferritin reconstituted without phosphate. Such results indicate that ferritin mineralization occurs in the plastid, where the phosphate content is higher; a corollary is the existence of a plastid Fe uptake system to allow the concentration of Fe in the ferritin mineral.

Mechanism of manganese catalase peroxide disproportionation: determination of manganese oxidation states during turnover

X-ray absorption near-edge structure (XANES) spectroscopy has been used to determine the oxidation state composition of the Mn site in Mn catalase under turnover conditions. The XANES data are consistent with parallel assignments based on electron paramagnetic resonance (EPR). However, a major advantage of the XANES assignments is that they permit the direct determination of the average oxidation states for derivatives that are EPR silent. In agreement with earlier work [Khangulov, S. V., Barynin, V. V., & Antonyuk-Barynina, S. V. (1990) Biochim. Biophys. Acta 1020, 25-33], these data show that the binuclear Mn site is reduced to Mn(II)/Mn(II) when peroxide is added in the presence of halide inhibitors. In addition, the present data provide the first direct evidence that the reduced enzyme is oxidized if peroxide is added in the absence of inhibitors. Under turnover conditions, the enzyme contains approximately a 2:1 ratio of Mn(II) and Mn(III). Similar results are obtained following incubation with dioxygen. These results are consistent with a Mn(II)/Mn(II)<==>Mn(III)/Mn(III) catalytic cycle and demonstrate that halide inhibition involves trapping the enzyme in the reduced state.

Crystallization and structural analysis of bullfrog red cell L-subunit ferritins

Ferritin is a 24 subunit protein that controls biomineralization of iron in animals, bacteria, and plants. Rates of mineralization vary among members of the ferritin family, particularly between L and H type subunits of animal ferritins which are differentially expressed in various cell types. To examine ferritin from a highly differentiated cell type and to clarify the relationship between ferritin structure and function, bullfrog red cell L ferritin has been cloned, overexpressed in E. coli, and crystallized under two conditions. Crystals were obtained at high ionic strength in the presence of MnCl2 at a concentration comparable to that of the protein and in the presence of MgCl2 at a concentration much higher than that of the protein. Under both crystallization conditions, the crystals are tetragonal bipyramids in the space group F432 with unit cell dimensions a = b = c = 182 +/- 0.5 A. Crystals obtained in the presence of manganese and ammonium sulfate diffract to 1.9 A, while those obtained in the presence of magnesium and sodium tartrate diffract to 1.6 A. Isomorphous crystals have been obtained under similar conditions for a site-directed mutant with a reduced mineralization rate in which Glu-57, -58, -59, and -61 are all replaced by Ala. The structure of wild type L-subunit with magnesium has been solved by molecular replacement using the calcium salt of human liver H subunit (Lawson et al., Nature (London) 349:541-544, 1991) as the model. The crystallographic R factor for the 6-2.2 A shell is 0.21. The overall fold of human H and bullfrog L ferritins is similar with an rms difference in backbone atomic positions of 0.97 A. The largest structural differences occur in the D helix and the loop connecting the D and E helices of the four helix bundle. Because red cell L ferritin and liver H ferritin show differences in both rates of mineralization and three-dimensional structure, more detailed comparisons of these structures are likely to shed new light on the relationship between conformation and function.

Formation of iron(III)-tyrosinate is the fastest reaction observed in ferritin

Rapid mineralization of ferritin, characteristic of protein with H-type subunits, coincides with formation of a specific Fe(III)-tyrosinate complex. The pseudo-first-order rate constant for Fe(II) oxidation by H-subunit-type ferritin has now been shown to be 700-900 times greater than any previously reported for ferritin; kox = 1000 s-1 for formation of the specific Fe(III)-tyrosinate complex (A550nm) or formation of less defined Fe(III)-oxo multinuclear complexes (A420nm). Formation of multinuclear Fe(III)-oxo complexes and O2 consumption were biphasic. In the first phase, up to 50 Fe atoms/ferritin molecule were rapidly oxidized, accompanied by formation of the Fe(III)-tyrosinate complex; saturation of the sites which formed the Fe(III)-tyrosinate complex also required 50 Fe/ferritin molecule. The sigmoidal shape of the curve obtained by plotting the initial rate of oxidation during the rapid phase of mineralization versus added [Fe(II)] suggested a more complex reaction pathway of ferroxidation than previously described. During the second phase of mineralization, Fe(III)-tyrosinate decreased, but multinuclear Fe(III)-oxo complexes and O2 consumption continued to increase at a slower rate. Recovery of the rapid oxidation pathway paralleled recovery of the site for Fe(III)-tyrosinate formation; full regeneration of the Fe(III)-tyrosinate sites was gradual over a period of 12 h, as if the movement of Fe(III) along the biomineralization pathway in the protein was slow and was accompanied by conformational changes which affected the Fe(III)-tyrosinate site.(ABSTRACT TRUNCATED AT 250 WORDS)

Formation of an Fe(III)-tyrosinate complex during biomineralization of H-subunit ferritin

An iron(III)-tyrosinate complex was identified in ferritin by ultraviolet-visible and resonance Raman spectroscopies. Previously, a specific amino acid side chain coordinated to iron in ferritin was not known. Ferritin protein was overexpressed in Escherichia coli from complementary DNA sequences of bullfrog red cell ferritin. The purple iron(III)-tyrosinate intermediate that formed during the first stages of iron uptake was replaced by the amber multinuclear iron(III)-oxo complexes of fully mineralized ferritin. Only the H subunit formed detectable amounts of the iron(III)-tyrosinate complex, which may explain the faster rates of iron biomineralization in H- compared to L-type ferritin.

Inactivation and reactivation of manganese catalase: oxidation-state assignments using X-ray absorption spectroscopy

The oxidation states of the Mn atoms in three derivatives of Mn catalase have been characterized using a combination of X-ray absorption near-edge structure (XANES) and EPR spectroscopies. The as-isolated enzyme has an average oxidation state of Mn(III) and contains a Mn(III) form, together with a reduced Mn(II) form and a variable amount (10-25%) of a Mn(III)/Mn(IV) mixed-valence derivative. Treatment with NH2OH rapidly reduces the majority of the enzyme to a Mn(II) derivative with no loss of activity. Inactivation by treatment with NH2OH + H2O2 converts all of the enzyme to a mixed-valence Mn(III)/Mn(IV) form. The inactive, mixed-valence derivative can be completely reactivated by long-term (greater than 1 h) anaerobic incubation with NH2OH, giving a reduced Mn(II)/Mn(II) derivative. These data suggest a catalytic model in which the enzyme cycles between a reduced Mn(II)/Mn(II) state and an oxidized Mn(III)/Mn(III) state.