A new subclass of the zinc metalloproteases superfamily revealed by the solution structure of peptide deformylase.

Escherichia coli peptide deformylase, a member of the zinc metalloproteases family, is made up of an active core domain composed of 147 residues and of an additional and dispensable C-terminal tail of 21 residues. The three-dimensional structure of the catalytic core could be studied by NMR. 1H and 15N NMR resonances assignments were obtained by two-dimensional and three-dimensional heteronuclear spectroscopy. The structure could be calculated using a set of 1015 restraints for the 147 residues of the enzyme. The overall structure is composed of a series of antiparallel beta-strands which surround two perpendicular alpha-helices. The C-terminal helix contains the HEXXH motif, which is crucial for activity. This helical arrangement and the way the histidines bind the zinc ion clearly are structurally reminiscent of the other members of the metalloprotease family, such as thermolysin or metzincins. Nevertheless, the overall arrangement of secondary and tertiary structures of peptide deformylase and the positioning of its third zinc ligand (a cysteine) are quite different from those of the other members of the family. These discrepancies, together with several biochemical differences, lead us to propose that peptide deformylase is the first example of a new class of the zinc-metalloproteases family. Studies of the interaction of peptide deformylase with either an inhibitor of the reaction or a product of the catalysed reaction, Met-Ala-Ser, as well as comparisons with the structures of other enzymes of the family, have enabled us to delineate the area corresponding to their binding site. The structural basis of the specificity of recognition of the formyl group is discussed in the context of the protease superfamily.

Methionyl-tRNA synthetase zinc binding domain. Three-dimensional structure and homology with rubredoxin and gag retroviral proteins.

Methionyl-tRNA synthetase from Escherichia coli contains one tightly bound zinc atom per subunit. The region encompassing residues 138 to 163 of this enzyme is responsible for the metal binding. A 28-mer peptide corresponding to these residues was expressed in vivo and shown to contain approximately 1 mol of tightly bound Zn/mol of peptide. In this study, the three-dimensional solution structure of this peptide was solved by means of two-dimensional proton NMR spectroscopy. A total of 133 nuclear Overhauser effect distance constraints and 22 dihedral angle restraints were used for the calculations, using a hybrid distance-geometry-simulated annealing strategy. Excluding the first four residues, the resulting structure is well-defined (r.m.s.d. 0.71 A for backbone atoms) and composed of a series of four tight turns. The second and the fourth turns are composed of CXXC sequences which are structurally homologous to the NH-S turns found in the metal binding sites of gag retroviral proteins and rubredoxin. The solution structure of the zinc binding peptide shows significant discrepancies with the crystal structure of methionyl-tRNA synthetase.

Solution structure of the anticodon-binding domain of Escherichia coli lysyl-tRNA synthetase and studies of its interaction with tRNA(Lys).

A protein domain corresponding to residues 31 to 149 of the E. coli Lysyl-tRNA synthetase species corresponding to the lysS gene was expressed and 15N-labelled. 1H and 15N NMR resonance assignments for this domain were obtained by two-dimensional and three-dimensional homonuclear and heteronuclear spectroscopy. Using distance geometry and simulated annealing, a three-dimensional structure could be calculated using 701 NOE and 86 dihedral angle restraints. It is composed of a five-stranded antiparallel beta-barrel capped by three alpha-helices at its ends. This structure closely resembles that of the N-terminal domain of the other E. coli lysyl-tRNA synthetase species expressed from the lysU gene and is highly homologous to the fold observed for the corresponding region of aspartyl-tRNA synthetase. It is shown that the isolated N-terminal fragment of lysyl-tRNA synthetase can interact with tRNA(Lys) as well as with poly (U), which mimics the anticodon sequence. Amino acid residues involved in these interactions were identified and, in the case of poly-U, a number of specific protein-RNA contacts were characterized. Specific recognition of tRNA(Lys) involves a cluster of four structurally well-defined aromatic residues, anchored on the beta-strands, and basic residues located on the surrounding loops. This organization is reminiscent of other RNA binding proteins, such as the U1A small nuclear ribonucleoprotein.

Heteronuclear NMR studies of the interaction of tRNA(Lys)3 with HIV-1 nucleocapsid protein.

Reverse transcription of HIV-1 viral RNA uses human tRNA(Lys)3 as a primer. Recombinant tRNA(Lys)3 was previously overexpressed in Escherichia coli, 15N-labelled and purified for NMR studies. It was shown to be functional for priming of HIV-1 reverse transcription. Using heteronuclear 2D and 3D NMR, we have been able to assign almost all the imino groups in the helical regions and involved in the tertiary base interactions of tRNA(Lys)3. This crucial step enabled us to address the question of the annealing mechanism of tRNA(Lys)3 by the nucleocapsid protein (NC) using heteronuclear NMR. Moreover, structural aspects of the tRNA(Lys)3/(12-53)NCp7 interaction have been characterised. The (12-53)NCp7 protein binds preferentially to the inside of the L-shape of the tRNA(Lys)3, on the acceptor and D stems, and at the level of the tertiary interactions between the D and T-psi-C loops. (12-53)NCp7 binding does not induce the melting of any single base-pair or unwinding of the tRNA(Lys)3 helical domains. Moreover, NMR provides a unique means to investigate the base-pairs that are destabilised by (12-53)NCp7 binding. Indeed, the measurements of the longitudinal relaxation time T1 and of the exchange time of the imino protons revealed two major regions sensitive to catalysis by the protein, namely the G6-U67 and T54(A58) pairs. It is interesting that for the biological role of the NC protein, these pairs could be the starting points of the tRNA melting required for the hybridisation to the viral RNA.

Lysine 335, part of the KMSKS signature sequence, plays a crucial role in the amino acid activation catalysed by the methionyl-tRNA synthetase from Escherichia coli.

The KMSKS pattern, conserved among several aminoacyl-tRNA synthetase sequences, was first recognized in the Escherichia coli methionyl-tRNA synthetase through affinity labelling with an oxidized reactive derivative of tRNA(Met)f. Upon complex formation, two lysine residues of the methionyl-tRNA synthetase (Lys61 and 335, the latter being part of the KMSKS sequence) could be crosslinked by the 3'-acceptor end of the oxidized tRNA. Identification of an equivalent reactive lysine residue at the active centre of tyrosyl-tRNA synthetase designated the KMSKS sequence as a putative component of the active site of methionyl-tRNA synthetase. To probe the functional role of the labelled lysine residue within the KMSKS pattern, two variants of methionyl-tRNA synthetase containing a glutamine residue at either position 61 or 335 were constructed by using site-directed mutagenesis. Substitution of Lys61 slightly affected the enzyme activity. In contrast, the enzyme activities were very sensitive to the substitution of Lys335 by Gln. Pre-steady-state analysis of methionyladenylate synthesis demonstrated that this substitution rendered the enzyme unable to stabilize the transition state complex in the methionine activation reaction. A similar effect was obtained upon substituting Lys335 by an alanine instead of a glutamine residue, thereby excluding an effect specific for the glutamine side-chain. Furthermore, the importance of the basic character of Lys335 was investigated by studying mutants with a glutamate or an arginine residue at this position. It is concluded that the N-6-amino group of Lys335 plays a crucial role in the activation of methionine, mainly by stabilizing the transient complex on the way to methionyladenylate, through interaction with the pyrophosphate moiety of bound ATP-Mg2+. We propose, therefore, that the KMSKS pattern in the structure of an aminoacyl-tRNA synthetase sequence represents a signature sequence characteristic of both the pyrophosphate subsite and the catalytic centre.

Three-dimensional structure of the lipoyl domain from Bacillus stearothermophilus pyruvate dehydrogenase multienzyme complex.

The structure of the lipoyl domain from the pyruvate dehydrogenase multienzyme complex of Bacillus stearothermophilus has been determined by means of nuclear magnetic resonance spectroscopy. A total of 452 nuclear Overhauser effect distance constraints and 76 dihedral angle restraints were employed as the input for the structure calculations, which were performed using a hybrid distance geometry-simulated annealing strategy and the programs DISGEO and X-PLOR. The overall structure of the lipoyl domain (residues 1 to 79 of the dihydrolipoamide acetyltransferase polypeptide chain) is that of a flattened eight-stranded beta-barrel folded around a core of well-defined hydrophobic residues. The lipoylation site, lysine 42, is located in the middle of a beta-turn, and the N and C-terminal residues of the domain are close together in adjacent beta-strands at the opposite end of the molecule. The polypeptide backbone exhibits a 2-fold axis of quasi-symmetry, with the C alpha atoms of residues 15 to 39 and 52 to 76 being almost superimposable on those of residues 52 to 76 and 15 to 39, respectively (root-mean-square deviation = 1.48 A). The amino acid residues at key positions in the structure are conserved among all the reported primary structures of lipoyl domains, suggesting that the domains all fold in a similar way.

Solution structure of nickel-peptide deformylase.

In the accompanying paper, we report that zinc is unlikely to be the co-factor supporting peptide deformylase activity in vivo. In contrast, nickel binding promotes full enzyme activity. The three-dimensional structure of the resulting nickel-containing peptide deformylase (catalytic domain, residues 1 to 147) was solved by NMR using a 13C-15N-doubly labelled protein sample. A set of 2261 restraints could be collected, with an average of 15.4 per amino acid. The resolution, which shows a good definition for the position of most side-chains, is greatly improved compared to that previously reported for the zinc-containing, inactive form. A comparison of the two stuctures indicates however that both share the same 3D organization. This shows that the nature of the bound metal is the primary determinant of the hydrolytic activity of this enzyme. Site-directed mutagenesis enabled us to determine the conserved residues of PDF involved in the structure of the active site. In particular, a buried arginine appears to be critical for the positioning of Cys90, one of the metal ligands. Furthermore, the 3D structure of peptide deformylase was compared to thermolysin and metzincins. Although the structural folds are very different, they all display a common structural motif involving an alpha-helix and a three-stranded beta-sheet. These conserved structural elements build a common scaffold which includes the active site, suggesting a common hydrolytic mechanism for these proteases. Finally, an invariant glycine shared by both PDF and metzincins enables us to extend the conserved motif from HEXXH to HEXXHXXG.

Mutations that alter initiation codon discrimination by Escherichia coli initiation factor IF3.

This work describes the isolation of mutations in infC, the structural gene for IF3, using different genetic screens. Among 21 mutants characterised, seven were shown to produce stable variant IF3 proteins unable to fully complement a strain carrying a chromosomal deletion of the infC gene. The mutants were also shown to be unable to normally discriminate against several non-canonical initiation codons such as AUU and ACG. The two mutants with the strongest complementation or discrimination defects carry changes in the C-terminal domain of IF3, which is responsible for the binding of the factor to the 30 S ribosomal subunit. We show that the first mutant has an expected decreased but the second an unexpected increased capacity to bind the 30 S subunit. The in vivo defects of the second mutant are explained by its capacity to bind unspecifically to other targets, as shown by its increased affinity for the 50 S subunit, which is normally not recognised by the factor. Interestingly, this mutant corresponds to a change of an acidic residue that might play a negative discriminatory role in preventing interactions with non-cognate RNAs, as has been reported for acidic residues of aminoacyl-tRNA synthetases shown to be involved in tRNA recognition.

Solution structure of the ribosome-binding domain of E. coli translation initiation factor IF3. Homology with the U1A protein of the eukaryotic spliceosome.

Initiation of translation in prokaryotes requires the formation of a complex between the messenger RNA, the 30 S ribosomal subunit and the initiator tRNA(fMet). Initiation factor IF3 binds to the 30 S ribosomal subunit and proof-reads the initiation complex, thereby ensuring the accuracy of this step. IF3 also plays a pleiotropic role in the regulation of translation, as a result of differential influences exerted on the levels of the initiation of translation of genes or groups of genes. IF3 is composed of two independent domain or roughly identical sizes. We have expressed and purified the C-terminal domain of E. coli IF3 and shown that it retains both the 30 S particle binding and 70 S ribosome dissociating activities of the native protein. We have obtained 1H and 15N NMR resonance assignments and its 3D solution structure was calculated using 551 restraints. It is composed of a mixed beta-sheet backed by two alpha-helices. It shows a striking resemblance to the U1A small nuclear ribonucleoprotein structure, which binds to the U1 snRNA in the eukaryotic spliceosome. This suggests a convergent evolution process for these two proteins that are associated with ribonucleoproteic complexes.

Heteronuclear NMR studies of E. coli translation initiation factor IF3. Evidence that the inter-domain region is disordered in solution.

Initiation factor IF3 from Escherichia coli plays a critical role in the selection of the correct initiation codon. This protein is composed of two domains, connected by a lysin-rich hydrophilic linker. The conformation of native IF3 was investigated by heteronuclear NMR spectroscopy. The two domains are independent and show little or no interaction. Heteronuclear relaxation studies of a sample selectively labelled on lysine residues demonstrates that the inter-domain linker is highly flexible, exhibiting increased 15N T2 values and negative 1H[15N] nuclear Overhause effects over a length of at least eight residues. Analysis of the rotational correlation times further shows that the motions of the two domains are most likely uncorrelated. The inter-domain linker thus displays almost totally unrestricted motions. Accordingly, the amide protons in the central region are shown to be in fast exchange with water. Such a high degree of flexibility of the inter-domain linker might be required for IF3 domains to interact with distant regions of the ribosome.

Expression in Escherichia coli of a sub-gene encoding the lipoyl domain of the pyruvate dehydrogenase complex of Bacillus stearothermophilus.

A sub-gene encoding the lipoyl domain (residues 1-85) of the lipoate acetyltransferase chain of the pyruvate dehydrogenase complex of Bacillus stearothermophilus was over-expressed in Escherichia coli. Approx. 80% of the domain was unlipoylated but most of the remainder was correctly lipoylated on Lys-42 and could be reductively acetylated by the B stearothermophilus enzyme complex. A small proportion (approx. 4%) of the domain carried an aberrant substituent, possibly an octanoyl group, on Lys-42. The 400 MHz 1H NMR spectra of the lipoylated and unlipoylated domains were essentially identical and closely resembled that of the native lipoyl domain.

The C-terminal domain of peptide deformylase is disordered and dispensable for activity.

Upon trypsinolysis, the 18 C-terminal residues of Escherichia coli peptide deformylase were removed but the resulting form exhibited full activity. Moreover, a mutant fms gene encoding the first 145 out of the 168 residues of the enzyme was able to complement a fms(Ts) strain and exhibited full activity. Upon progressive truncation up to residue 139, both activity and stability decreased up to complete inactivation. Mutagenesis of residues of the 138-145 region highlights the importance of Leu-141 and Phe-142. N-Terminal deletions were also carried out. Beyond two residues off, the enzyme showed a dramatic instability. Finally, NMR and thermostability studies of the full-length enzyme and comparison to the 1-147 form strongly suggest that the dispensable residues are disordered in solution.

The N-terminal half of initiation factor IF3 is folded as a stable independent domain.

Initiation factor IF3 plays an essential role in the initiation of protein translation by binding to the 30S ribosomal subunit and selecting a proper tRNA(fMet)/initiation codon complex. The domain structure of IF3 from Escherichia coli has been investigated by limited proteolysis followed by mass spectrometry and protein sequencing of the resulting peptides. This analysis revealed a highly segmented structure with two independent domains connected by a charged linker peptide, highly susceptible to proteolytic cleavage. The N-terminal domain is very stable and comparison of its 2-D NMR spectrum with that of intact IF3 revealed that it retains its three-dimensional fold.

Open reading frames in the control regions of the phenylalanyl-tRNA synthetase operon of E. coli.

The pheST operon codes for the two subunits of phenylalanyl-tRNA synthetase and it expression is controlled by attenuation in a way similar to many amino acid biosynthetic operons. The nucleotide sequence of the control regions of the operon indicates the presence of several open reading frames besides that of the leader peptide. One of these open reading frames, called the alternative leader peptide, starts at about the same place as the leader peptide and ends after the terminator of the attenuator. Another open reading frame, called the terminator peptide, starts after the terminator and covers about half the distance to pheS, the first structural gene of the operon. The present report shows that, in fact, the only open reading frame to be translated efficiently is the leader peptide itself. The alternative leader peptide and the terminator peptide are both translated at a negligible rate.

Genetic engineering of methionyl-tRNA synthetase: in vitro regeneration of an active synthetase by proteolytic cleavage of a methionyl-tRNA synthetase--beta-galactosidase chimeric protein.

The construction of a family of plasmids carrying derivatives of metG, the gene for E. coli methionyl-tRNA synthetase, is described. These plasmids allow expression of native or truncated forms of the enzyme and easy purification of the products. To facilitate the characterization of modified enzymes with very low catalytic activity, a specialized vector was constructed, in which metG was fused in frame with lacZ, the gene for beta-galactosidase. This plasmid expresses a methionyl-tRNA synthetase-beta-galactosidase chimeric protein, which is shown to carry the activities of both enzymes. This hybrid can be purified in a single step of affinity chromatography for beta-galactosidase. The methionyl-tRNA synthetase moiety can be regenerated by mild proteolysis, thus providing a simple method for purifying and studying mutated proteins.

Molecular recognition governing the initiation of translation in Escherichia coli. A review.

Selection of the proper start codon for the synthesis of a polypeptide by the Escherichia coli translation initiation apparatus involves several macromolecular components. These macromolecules interact in a specific and concerted manner to yield the translation initiation complex. This review focuses on recent data concerning the properties of the initiator tRNA and of enzymes and factors involved in the translation initiation process. The three initiation factors, as well as methionyl-tRNA synthetase and methionyl-tRNA(f)Met formyltransferase are described. In addition, the tRNA recognition properties of EF-Tu and peptidyl-tRNA hydrolase are considered. Finally, peptide deformylase and methionine aminopeptidase, which catalyze the amino terminal maturation of nascent polypeptides, can also be associated to the translation initiation process.

Methionyl-tRNA synthetase from E. coli--a review.

Methionyl-tRNA synthetase (MetRS) from E coli is a dimer composed of 2 identical subunits of Mr 76 kDa. A fully active monomeric fragment (64 kDa) could be obtained by mild proteolysis of the native dimer. Earlier studies reviewed in Blanquet et al (1979) have compared the catalytic mechanisms of native and trypsin-modified MetRS. Moreover, the truncated form of the enzyme was crystallized and its 3-D structure solved at low resolution. In the last few years, the availability of the corresponding metG gene has facilitated the development of studies using affinity labelling and site-directed mutagenesis techniques. In parallel, the 3-D structure has been solved at a resolution of 2.5 A. These convergent approaches have allowed significant progress in the understanding of the structure-function relationships of this enzyme, and, in particular, of the rules governing the recognition of tRNA.

Computer simulation of DNA ligation: determination of initial DNA concentrations favouring the formation of recombinant molecules.

A computer program was used to simulate the dynamic process of a ligation of DNA fragments. More specifically, the influence of the initial DNA fragments lengths and concentrations on the relative abundance of the various end-products was systematically investigated. Depending on the nature of the DNA extremities (asymmetric or symmetric, dephosphorylated or not), sets of initial conditions could be found that optimized the yield of active recombinant molecules. These results can be directly used to increase the efficiency of the ligation step, in particular for the construction of cDNA or genomic libraries.

MC-Fit: using Monte-Carlo methods to get accurate confidence limits on enzyme parameters.

A program is described for estimating enzymatic parameters from experimental data using Apple Macintosh computers. MC-Fit uses iterative least-square fitting and Monte-Carlo sampling to get accurate estimates of the confidence limits. This approach is more robust than the conventional covariance matrix estimation, especially in cases where experimental data is partially lacking or when the standard error on individual measurements is large. This happens quite often when analysing the properties of variant enzymes obtained by mutagenesis, as these can have severely impaired activities and reduced affinities for their substrates.