The interplay between X-ray crystallography, neutron diffraction, image reconstruction, organo-metallic chemistry and biochemistry in structural studies of ribosomes.

Crystals of ribosomes, their complexes with components of protein biosynthesis, their natural, mutated and modified subunits, have been subjected to X-ray and neutron crystallographic analyses. Electron microscopy and 3-dimensional image reconstruction, supported by biochemistry, genetic, functional and organo-metallic studies were employed for facilitating phasing of the crystallographic data. For example, a monofunctional multi heavy-atom cluster (undecagold) was designed for covalent and quantitative binding to ribosomes. The modified particles were crystallized isomorphously with the native ones. Their difference-Patterson maps contain indications for the usefulness of these derivatives for subsequent phasing. Models of the ribosome and its large subunit were reconstructed from tilt series of 2-dimensional sheets. The comparison of the various reconstructed images enabled an initial assessment of the reliability of these models and led to tentative assignments of several functional features. These include the presumed sites for binding mRNA and for codon-anticodon interactions, the path taken by the nascent protein chain and the mode for tRNA binding to ribosomes. These assignments assisted in the design of biologically meaningful crystal systems. The reconstructed models are being used to identify structural features in initial density maps derived from X-ray and neutron diffraction data.

Isolation and characterization of Bacillus stearothermophilus 30S and 50S ribosomal protein mutations.

Bacillus stearothermophilus mutations which confer resistance to or dependence on a variety of ribosome-targeted antibiotics have been isolated. Many of these mutations produce ribosomal proteins with altered mobilities in a two-dimensional gel electrophoresis system. This collection of altered thermophilic ribosomal proteins will be useful in examining ribosomal structure and function.

Overexpression of the methanococcal ribosomal protein L12 in Escherichia coli and its incorporation into halobacterial 50 S subunits yielding active ribosomes.

The gene for the ribosomal L12 protein from the archaebacterium Methanococcus vannielii was cloned into the expression vector pKK223-3. The protein was overexpressed and remained stable in Escherichia coli XL1 cells. Purification yielded a protein with the same amino acid composition and sequence as in Methanococcus but it was acetylated at the N terminus as in the case with the homologous protein of E. coli. The in vivo incorporation of the overexpressed protein into the E. coli ribosomes was not observed. The overexpressed M. vannielii protein MvaL12e was incorporated into halobacterial ribosomes, thereby displacing the corresponding halobacterial L12 protein. Intact 70 S ribosomes were reconstituted from halobacterial 50 S subunits carrying the MvaL12e protein. These ribosomes were as active as native halobacterial ribosomes in a poly(U) assay. On the other hand, our attempts to incorporate L12 proteins from Bacillus stearothermophilus and E. coli into halobacterial ribosomes were not successful. These results support the conclusion which is based on primary sequence and predicted secondary structure comparisons that there exist two distinct L12 protein families, namely the eubacterial L12 protein family and the eukaryotic/archaebacterial L12 protein family.

Studies on the accessibility of nascent non-helical peptide chains on the ribosome.

The accessibility of nascent polypeptides with special structural elements to the ribosome was investigated. Poly(C), poly(C, U) and poly(C, A) mRNAs were translated by E. coli ribosomes in vitro. The resulting peptides which were rich in prolines, remained on the ribosomal particles or were released after addition of puromycin. A protease from Aspergillus oryzae hydrolyzed the released peptides rapidly, whereas the degradation of the unreleased ones was only slightly affected. This result shows that the nascent peptides were protected against proteolytic attack by the ribosomal particles. Interestingly, the protease completely degraded the 30S particles whereas the 50S ones remained intact, even after prolonged incubation.

Characterization of crystals of small ribosomal subunits.

Crystals of intact small ribosomal subunits from Thermus thermophilus have been obtained from functionally active particles. The crystals (P42(1)2, 407 A x 407 A x 171 A) are suitable for X-ray crystallography analysis to 9.9 A using synchrotron radiation at cryotemperature. Crystallographic data from native and a potential heavy-atom derivative have been collected.

Two-dimensional crystalline sheets of Bacillus stearothermophilus 50S ribosomal subunits containing a nascent polypeptide chain.

Polylysine chains were synthesized on Bacillus stearothermophilus ribosomes in a poly(A)-programmed in vitro system. After separation of the ribosomal subunits by sucrose gradient centrifugation, the polylysine chains (in contrast to the polyphenylalanine chains synthesized in a poly(U) system) reproducibly remained attached to the large ribosomal subunit. It was possible to produce two-dimensional crystalline sheets from the large ribosomal subunits containing the polylysine chains. These sheets are an essential prerequisite for three-dimensional reconstruction studies aiming to show that the tunnel in the large ribosomal subunit provides a path for the nascent polypeptide chain.

The growth of ordered two-dimensional sheets of ribosomal particles from salt-alcohol mixtures.

A procedure for the in vitro growth of well-ordered two-dimensional sheets from ribosomal particles using salts and salt-alcohol mixtures has been developed. Employing this procedure, ordered two-dimensional sheets of the wild type as well as of mutated 50 S ribosomal subunits from Bacillus stearothermophilus can readily be obtained. These sheets, stained with uranyl acetate or gold-thioglucose, are suitable for three-dimensional image reconstruction. They consist of relatively small unit cells with dimensions of 160 +/- 15 and 365 +/- 20 A. Diffraction patterns of electron micrographs of these sheets contain features to 25 A resolution.

Reconstitution and crystallisation experiments with isolated split proteins from Bacillus stearothermophilus ribosomes.

Six proteins (B-L1, B-L6, B-L10, B-L11, B-L12 and B-L16) were removed from 50S ribosomal subunits of Bacillus stearothermophilus by treatment with ethanol and ammonium chloride. The proteins were isolated in a pure form, and one of them (B-L6) was crystallized. Five of the six proteins (in various combinations) were added back to the core particles, resulting in 50S subunits lacking one protein. The biological activities of these ribosomal particles as determined in the poly(U)-system varied over a wide range, depending on the protein which was omitted. The particles lacking one protein provide useful tools for heavy-atom derivation necessary for our crystallographic studies on the 50S subunits of Bacillus stearothermophilus.

Three-dimensional crystals of ribosomes and their subunits from eu- and archaebacteria.

Ordered three-dimensional crystals of 70S ribosomes as well as of 30S and 50S ribosomal subunits from various bacteria (E. coli, Bacillus stearothermophilus, Thermus thermophilus and Halobacterium marismortui) have been grown by vapour diffusion in hanging drops using mono- and polyalcohols. A new compact crystal form of 50S subunits has been obtained, and it is suitable for crystallographic studies at medium resolution. In addition, from one crystal form large crystals could be grown in X-ray capillaries. In all cases the crystals were obtained from functionally active ribosomal particles, and the particles from dissolved crystals retained their integrity and biological activity.

Three-dimensional image reconstruction from ordered arrays of 70S ribosomes.

A better understanding of the molecular mechanism of protein biosynthesis still awaits a reliable model for the ribosomal particle. We describe here the application of a diffraction technique, namely three-dimensional image reconstruction from two-dimensional sheets of 70S ribosomes from Bacillus stearothermophilus at 47 A resolution. The three-dimensional model obtained by these studies shows clearly the two subunits, the contact points between them, an empty space large enough to accommodate the components of protein biosynthesis, the location of regions rich in RNA and a possible binding site for mRNA. The tunnel within the 50S particle which may provide the path taken by the nascent polypeptide chain in partially resolved.

The growth of ordered two-dimensional sheets of 70 S ribosomes from Bacillus stearothermophilus.

Well ordered two-dimensional sheets of intact 70 S ribosomes from Bacillus stearothermophilus have been obtained in vitro using salt-alcohol mixtures. These sheets consist of relatively small unit cells with dimensions of 200 +/- 20 A and 400 +/- 30 A. Diffraction patterns of electron micrographs of these sheets stained with uranyl acetate contain features to 42 A resolution.

A compact three-dimensional crystal form of the large ribosomal subunit from Bacillus stearothermophilus.

A new form of well-ordered three-dimensional crystals of intact 50 S ribosomal subunits from Bacillus stearothermophilus have been obtained. Electron micrographs of positively stained sections of these crystals revealed that the ribosomal particles are packed closely. The cell parameters have been determined. Representative electron micrographs and their computed contoured filtered images are shown.

Purification and characterization of demolybdo nitrate reductase (NADH-cytochrome c oxidoreductase) of Chlorella vulgaris.

Chlorella vulgaris was cultured on an ammonia-mineral salts medium until the nitrate reductase content reached a minimal level. These ammonia-grown cells were then induced by nitrate in the absence of molybdenum and of tungsten. A demolybdo nitrate reductase developed and reached high levels. This protein contained very little nitrate-reducing capacity, but had the full cytochrome c-reducing capacity of normal nitrate reductase. It was purified to homogeneity by the same procedures previously developed for the purification of nitrate reductase. The purified enzyme contained 1 molecule of heme and 1 molecule of FAD/subunit, but no detectable molybdenum or tungsten. This cytochrome c reductase was completely inhibited by antibodies raised against purified nitrate reductase of Chlorella. Mixtures prepared from normal nitrate reductase and the demolybdoenzyme could not be resolved by disc gel electrophoresis or by centrifugation in a density gradient. By a two-step enzyme induction (1, incubation with nitrate in absence of Mo; 2, incubation with Mo in absence of nitrate) the process of nitrate reductase synthesis could be cleanly separated from growth into two steps: Step 1, induction of cytochrome c reductase, was completely inhibited by cycloheximide. Step 2 was unaffected by cycloheximide, and most of the nitrate reductase synthesized accumulated in the form of the reversibly inactivated HCN complex of the enzyme.

The formation of hydrogen cyanide from histidine in the presence of amino acid oxidase and peroxidase.

Conditions were sought to increase the yield of HCN from L-histidine incubated with L-amino acid oxidase (L-amino acid:oxygen oxidoreductase (deaminating), EC 1.4.3.2) from snake venom, and horseradish peroxidase (donor:hydrogen-peroxide oxidoreductase, EC 1.11.1.7). Small amounts of histidine and high buffer concentrations favored high HCN yields, which reached a maximum of 72%. Imidazole 4-aldehyde and imidazole 4-carboxylic acid were identified among the reaction products, together with CO2, NH3, H2O2 and imidazole acetic acid. The CO2 formed was equal to the histidine oxidized, and to the sum of NH3 plus HCN formed. The production of HCN was associated with an increased O2 uptake, which was established from the beginning of the reaction, with no apparent lag and ranged from 1.2 to 1.6 mumol extra O2 taken up/mumol HCN formed. The system was inhibited by catalase, but added superoxide dismutase caused a small stimulation of both HCN production and O2 consumption, and a larger stimulation of H2O2 accumulation. Added hydroxylamine was cooxidized to nitrite in an amount equimolar with the HCN formed. This nitrite formation was inhibited by superoxide dismutase. The facts could be interpreted in terms of superoxide anion formation during the HCN-producing reaction. cytochrome c, heme, or ferricyanide could be substituted for peroxidase, but were less effective. The initial rates of HCN formation from phenylalanine, tyrosine and tryptophan were higher, but the eventual yields of HCN from these amino acids were lower than those from histidine.

Nitrate reductase of Chlorella fusca: Partial purification, cytochrome content and presence of HCN after in vivo inactivation.

An inactivated nitrate reductase (EC 1.6.6.1) formed in vivo by the green alga Chlorella fusca Shihira and Kraus is shown to be a cyanide complex. The partially purified inactive enzyme releases 0.048 nmol of HCN per unit of enzyme activated. This compares with 0.066 nmol of HCN liberated in similar previous measurements with the inactivated enzyme from Chlorella vulgaris. The nitrate reductase from C. fusca has been purified to a level of 67 µmol nitrate reduced per min per mg enzyme. It contains a cytochrome b557, at a level 1.9-fold higher per unit of active enzyme, than the nitrate reductase from C. vulgaris.

Cyanide formation from histidine in Chlorella. A general reaction of aromatic amino acids catalyzed by amino acid oxidase systems.

The formation of HCN from D-histidine in Chlorella vulgaris extracts is shown to be due to the combined action of a soluble protein and a particulate component. Either horse-radish peroxidase (EC 1.11.1.7) or a metal ion with redox properties can be substituted for the particulate component. Ions of manganese and vanadium are especially effective, as are o-phenanthroline complexes of iron. Cobalt ions are less active. The D-amino acid oxidase (EC 1.4.3.3) from kidney and the L-amino acid oxidase (EC 1.4.3.2) from snake venom likewise cause HCN production from histidine when supplemented with the particulate preparation from Chlorella or with peroxidase or with a redox metal ion. The stereospecificity of the amino acid oxidase determines which of the two stereoisomers of histidine is active as an HCN precursor. Though histidine is the best substrate for HCN production, other naturally occurring aromatic amino acids (viz. tyrosine, phenylalanine and tryptophan) can also serve as HCN precursors with these enzyme systems. The relative effectiveness of each substrate varies with the amino acid oxidase enzyme and with the supplement. With respect to this latter property, the particulate preparation from Chlorella behaves more like a metal ion than like peroxidase.

Cyanide formation in preparations from Chlorella vulgaris Beijerinck: Effect of sonication and amygdalin addition.

A critical evaluation of a method for recovering HCN from cell extracts is presented. Since crude extracts often bind or metabolize HCN extensively, the HCN recovered by distillation at room temperature represents only the difference between production and consumption. Sonication leads to HCN release from the alga, Chlorella vulgaris Beijerinck. Illumination of extracts at high light intensity in oxygen, with added Mn(2+) and peroxidase, also stimulates HCN production. In both processes, the HCN is probably formed by oxidation of nitrogenous precursors. Chlorella extracts cause formation of HCN from added amygdalin. No evidence was found, however, for the presence of cyanogenic glycosides in the algae.

Cyanide formation in preparations from Chlorella and New Zealand spinach leaves: Effect of added amino acids.

As part of an effort to identify the natural precursor(s) of HCN in the alga Chlorella vulgaris Beijerinck, and in leaves of New Zealand spinach (Tetragonia expansa, Murr.), HCN release was measured after addition of various amino acids to illuminated algal extracts and grana preparations. Histidine is particularly effective as an HCN precursor, both with Chlorella extracts and leaf grana. With the algal extracts, D-histidine is about ten times more effective than L-histidine and histamine, whereas the two isomers (and histamine) are about equally effective with leaf grana. In the presence of leaf grana plus added Mn(2+) and peroxidase, L-tyrosine and L-cysteine like-wise cause HCN formation; but these amino acids cause little or no HCN formation in the presence of Chlorella extracts. A stimulation of HCN production by L-histidine was observed with intact Chlorella cells. Because of the limitations of the assay method, the possibility can not be excluded that other substances than histidine may also lead to HCN generation in Chlorella vulgaris, but the results show that histidine has an important role in HCN generation by this species.

Reversible inactivation of nitrate reductase in Chlorella vulgaris in vivo.

The NADH-nitrate oxidoreductase of Chlorella vulgaris has an inactive form which has previously been shown to be a cyanide complex of the reduced enzyme. This inactive enzyme can be reactivated by treatment with ferricyanide in vitro. In the present study, the activation state of the enzyme was determined after different prior in vivo programs involving environmental variations. Oxygen, nitrate, light and CO2 all affect the in vivo inactivation of the enzyme in an interdependent manner. In general, the inactivation is stimulated by O2 and inhibited by nitrate and CO2. Light may stimulate or inhibit, depending on conditions. Thus, the effects of CO2 and nitrate (inhibition of reversible inactivation) are clearly manifested only in the light. In contrast, light stimulates the inactivation in the presence of oxygen and the absence of CO2 and nitrate. Since the inactivation of the enzyme requires HCN and NADH, and it is improbable that O2 stimulates NADH formation, it is reasonable to conclude that HCN is formed as the result of an oxidation reaction (which is stimulated by light). The formation of HCN is probably stimulated by Mn(2+), since the formation of reversibly-inactivated enzyme is impaired in Mn(2+)-deficient cells. The prevention of enzyme inactivation by nitrate in vivo is in keeping with previous in vitro results showing that nitrate prevents inactivation by maintaining the enzyme in the oxidized form. A stimulation of nitrate uptake by CO2 and light could account for the effect of CO2 (prevention of inactivation) which is seen mainly in the presence of nitrate and light. Ammonia added in the presence of nitrate has the same effect on the enzyme as removing nitrate (promotion of reversible inactivation). Ammonia added in the absence of nitrate has little extra effect. It is therefore likely that ammonia acts by preventing nitrate uptake. The uncoupler, carbonylcyanide-m-chloro-phenylhydrazone, causes enzyme inactivation because it acts as a good HCN precursor, particularly in the light. Nitrite, arsenate and dinitrophenol cause an enzyme inactivation which can not be reversed by ferricyanide in crude extracts. This suggests that there are at least two different ways in which the enzyme can be inactivated rather rapidly in vivo.