Variation of the intercalating proline in artificial peptides mimicking the DNA binding and bending IHF protein.

The integration host factor (IHF) is a protein which sequence specifically induces a bend of double-stranded DNA by more than 160°. Based on IHF as lead structure, a peptide mimic was introduced resembling the positively charged body of the protein by a lysine dendrimer and the minor groove recognition loop by a cyclopeptide. The proline located close to the tip of the recognition loop intercalates between the base pair plane. It was modified in order to evaluate the influence of the side chain residue with respect to size (1,2,3,4-tetrahydroisoquinoline-3-carboxylic acid), aromaticity (phenylalanine), conformation of the five-membered ring [(4R)-fluoroproline, (4S)-fluoroproline, 3,4-dehydroproline], and the peptide backbone conformation (α-methylproline) on binding dsDNA and bending the double strand. Binding and bending studies were carried out by fluorescence resonance energy transfer experiments and gel electrophoresis using DNA sequences prepared by PCR with the IHF binding site in central or terminal position. Whereas aromatic residues and α-methylproline were not tolerated as proline substitute, incorporation of (4S)-fluoroproline and 3,4-dehydroproline provided enhanced binding.

1.7 A structure of the stabilized REIv mutant T39K. Application of local NCS restraints.

The X-ray structure of the T39K mutant of the variable domain of a human immunoglobulin kappa light chain has been determined at room temperature to 1.7 A resolution with a conventional R factor of 0. 182. T39K crystallizes in the triclinic space group P1 [a = 35.4 (1), b = 40.1 (1), c = 43.1 (1) A, alpha = 66.9 (1), beta = 85.4 (1), gamma = 73.8 (1) degrees ]. The unit-cell contains two monomers, related by a non-crystallographic twofold axis. The use of a novel type of local non-crystallographic symmetry restraints on related isotropic displacement parameters and 1-4 distances as incorporated in the refinement program SHELXL improves the model and quality of the maps, but local differences between both monomers in areas subject to different packing contacts can still be observed. 12 overall anisotropic scaling parameters were refined. These may have compensated for the difficulties in accurately scaling single rotation axis image plate data from a triclinic crystal, because of the scarcity of common equivalent reflections. The final model has been used to perform a number of tests on anisotropic scaling, non-crystallographic symmetry, anisotropic refinement, determination of standard uncertainties and bulk solvent correction. It is remarkable that removal of the NCS restraints from the final model caused Rfree to increase. These tests clarify the strategies for optimum use of SHELXL for refinement at medium as opposed to atomic resolution.

ToxR co-operative interactions are not modulated by environmental conditions or periplasmic domain conformation.

ToxR is a transmembrane regulatory protein that controls virulence gene expression in Vibrio cholerae. Previous experiments using lambda repressor-ToxR chimeric proteins and a lambda repressor-controlled reporter system (OR1 PR-lacZY) established that ToxR sequences can effectively dimerize the amino-terminal domain of lambda repressor in Escherichia coli. However, in E. coli, ToxR does not respond to environmental signals that control virulence gene expression in V. cholerae. Here, we report the results of experiments designed to test whether environmental signals that modulate virulence gene expression in V. cholerae also modulate a monomer to dimerization transition of lambda-ToxR chimeras. When the OR1 PR-lacZY reporter fusion and chimeric proteins were transferred to V. cholerae, we unexpectedly found that lambda-ToxR chimeras did not dimerize significantly. Interestingly, experiments evaluating the ability of lambda-ToxR proteins to form tetramers in E. coli suggested that lambda-ToxR dimers could act co-operatively. Using a redesigned reporter system containing multiple lambda operator sites (OR1 OR2 OR3 PR-lacZY), we found that lambda-ToxR could dimerize quite efficiently in V. cholerae. These data imply that multiple DNA binding sites might enhance the ability of ToxR to dimerize in V. cholerae and suggest that ToxR dimers might be capable of co-operative interactions. However, we falled to correlate a monomer-dimer transition of the lambda-ToxR chimeras with changes in virulence gene expression in response to environmental signals in V. cholerae. Finally, because of conflicting results in the literature, the importance of membrane localization of ToxR and dimerization of the ToxR periplasmic domain was re-evaluated. This was accomplished by measuring the ability of various chimeric proteins to activate toxin gene expression in both E. coli and V. cholerae. These assays suggest that, in V. cholerae, deletion of the transmembrane domain has a profound effect on ToxR activity, although it is not an absolute requirement when ToxR is dimerized by a heterologous domain. In addition, we noted differences in chimeric protein activity when expressed in E. coli and V. cholerae. A construct substituting the monomeric MalE domain for the periplasmic domain of ToxR was unable to activate a ctx::lacZ reporter fusion in E. coli. Although the addition of leucine zipper sequences to this construct resulted in enhanced activity of the chimera in E. coli, both chimeras were able to produce wild-type levels of toxin in V. cholerae. These data support the notion that dimerization of ToxR stimulates its activity as a transcriptional activator in E. coli. In V. cholerae, however, we present data that do not demonstrate a correlation between dimerization of the periplasmic domain and ToxR activity.

The Escherichia coli MutL protein stimulates binding of Vsr and MutS to heteroduplex DNA.

Vsr DNA mismatch endonuclease is the key enzyme of very short patch (VSP) DNA mismatch repair and nicks the T-containing strand at the site of a T-G mismatch in a sequence-dependent manner. MutS is part of the mutHLS repair system and binds to diverse mismatches in DNA. The function of the mutL gene product is currently unclear but mutations in the gene abolish mutHLS -dependent repair. The absence of MutL severely reduces VSP repair but does not abolish it. Purified MutL appears to act catalytically to bind Vsr to its substrate; one-hundredth of an equivalent of MutL is sufficient to bring about a significant effect. MutL enhances binding of MutS to its substrate 6-fold but does so in a stoichiometric manner. Mutational studies indicate that the MutL interaction region lies within the N-terminal 330 amino acids and that the MutL multimerization region is at the C-terminal end. MutL mutant monomeric forms can stimulate MutS binding.

X-ray crystallography reveals stringent conservation of protein fold after removal of the only disulfide bridge from a stabilized immunoglobulin variable domain.

BACKGROUND: Immunoglobulin domains owe a crucial fraction of their conformational stability to an invariant central disulfide bridge, the closure of which requires oxidation. Under the reducing conditions prevailing in cell cytoplasm, accumulation of soluble immunoglobulin is prohibited by its inability to acquire and maintain the native conformation. Previously, we have shown that disulfide-free immunoglobulins can be produced in Escherichia coli and purified from cytoplasmic extracts. RESULTS: Immunoglobulin REIv is the variable domain of a human kappa light chain. The disulfide-free variant REIv-C23V/Y32H was crystallized and its structure analyzed by X-ray crystallography (2.8 A resolution). The conformation of the variant is nearly identical to that of the wild-type protein and the conformationally stabilized variant REIv-T39K. This constitutes the first crystal structure of an immunoglobulin fragment without a disulfide bridge. The lack of the disulfide bridge produces no obvious local change in structure (compared with the wild type), whereas the Y32H mutation allows the formation of an additional hydrogen bond. There is a further change in the structure that is seen in the dimer in which Tyr49 has flipped out of the dimer interface in the mutant. CONCLUSIONS: Immunoglobulin derivatives without a central disulfide bridge but with stringently conserved wild-type conformation can be constructed in a practical two-step approach. First, the protein is endowed with additional folding stability by the introduction of one or more stabilizing amino acid exchanges; second, the disulfide bridge is destroyed by substitution of one of the two invariant cysteines. Such derivatives can be accumulated in soluble form in the cytoplasmic compartment of the E. coli cell. Higher protein yields and evolutionary refinement of catalytic antibodies by genetic complementation are among the possible advantages.

Dimerisation of the glycophorin A transmembrane segment in membranes probed with the ToxR transcription activator.

Specific interactions between membrane spanning polypeptide segments are important for folding and oligomerisation of integral membrane proteins. Previously the dimerisation of glycophorin A has been shown to depend on interactions between its transmembrane segment by studying chimeric proteins in detergent solution. Here, we examined dimerisation of the glycophorin A transmembrane segment in a natural membrane employing the ToxR transcription activator from Vibrio cholerae. The ToxR protein is integral to the bacterial inner membrane and its activity requires a dimeric state. Therefore, the ToxR protein is suited to monitor quantitative homophilic interactions. We replaced the ToxR transmembrane segment with parts of the glycophorin A transmembrane segment containing the amino acid motif LIxxGVxxGVxxT previously shown to be sufficient for dimerisation in detergent solution. Expression of these chimeric proteins in an indicator strain resulted in strong transcription activation. This is indicative of efficient dimerisation mediated by the glycophorin transmembrane segment inserted into the inner membrane. Analysis of individual point mutants revealed that at least four residues out of this motif are critical for dimer formation in membranes. However, dimerisation of the glycophorin A transmembrane segment appears to be less sensitive to mutations when localised within a natural lipid bilayer compared to measurements in detergent solution. This may be related to a slightly altered structure of the dimer and/or to a higher local concentration and preorientation of the interacting molecules in a membrane. This makes the ToxR system well suited for probing low-affinity interactions between the transmembrane segments of other proteins.

Statistical evidence for a biochemical pathway of natural, sequence-targeted G/C to C/G transversion mutagenesis in Haemophilus influenzae Rd.

Markov chain analysis of the Haemophilus influenzae Rd genome reveals striking under-representation of three palindromic tetranucleotide strings (CCGG, GGCC and CATG), accompanied by over-representation of six tetranucleotide strings that are derived from the former by exchanging strand location of the two residues making up a G/C nucleotide pair at the terminal palindrome position. Constraints are outlined for a molecular model able to explain the phenomenon as the result of sequence-targeted, enzyme-driven G/C to C/G transversion mutagenesis. Possible participation in the process by components of known DNA mismatch repair or restriction/modification systems (in particular, cytosine methylation) is discussed. The effect widens the spectrum of enzyme-driven, specific mutagenesis beyond the formerly described C/G to T/A transition (VSP repair of Escherichia coli). Potential evolutionary benefits of enzymatic pathways of specific mutagenesis can be envisioned.

Counteracting the mutagenic effect of hydrolytic deamination of DNA 5-methylcytosine residues at high temperature: DNA mismatch N-glycosylase Mig.Mth of the thermophilic archaeon Methanobacterium thermoautotrophicum THF.

Spontaneous hydrolytic deamination of DNA 5-methylcytosine residues gives rise to T/G mismatches which are pre-mutagenic lesions requiring DNA repair. For fundamental reasons, the significance of this and other processes lowering genetic fidelity must be accentuated at elevated temperatures, making thermophilic organisms attractive objects for studying how cells cope with thermal noise threatening the integrity of their genetic information. Gene mig of Methanobacterium thermoautotrophicum THF, an anaerobic archaeon with an optimal growth temperature of 65 degrees C, was isolated and its product (Mig.Mth; EC3.2.2-) shown to be a T/G-selective DNA thymine N-glycosylase with the properties required for counteracting the mutagenic effect of hydrolytic 5-meC deamination. The enzyme acts on T/G and U/G oppositions with similar efficiency; G/G, A/G, T/C and U/C are minor substrates; no other opposition of common nucleobases is attacked and no removal of U from single-stranded DNA is observed. Substrate preferences are modulated by sequence context. Together with the results presented here, one example of an enzyme directed against the hydrolytic deamination damage of 5-meC is known from each of the three phylogenetic kingdoms; entry into the repair pathway is glycosylytic in the eukaryotic and the archaeal case, whereas the eubacterial repair starts with an endonucleolytic DNA incision.

Contribution of the intramolecular disulfide bridge to the folding stability of REIv, the variable domain of a human immunoglobulin kappa light chain.

BACKGROUND: Immunoglobulin domains contain about 100 amino acid residues folded into two beta-sheets and stabilized in a sandwich by a conserved central disulfide bridge. Whether antibodies actually require disulfide bonds for stability has long been a matter of debate. The contribution made by the central disulfide bridge to the overall folding stability of the immunoglobulin REIv, the variable domain of a human kappa light chain, was investigated by introducing stabilizing amino acid replacements followed by removal of the disulfide bridge via chemical reduction or genetic substitution of the cysteine residues. RESULTS: Nine REIv variants were constructed by methods of protein engineering that have folding stabilities elevated relative wild-type REIv by (up to) 16.0 kJ mol-1. Eight of these variants can be cooperatively refolded after unfolding and chemical reduction of the disulfide bridge-in contrast to wildtype REIv. The stabilizing effect of one of these residue replacements (T39K) was rationalized by determining the structure of the respective REIv variant at 1.7 A. The loss of folding stability caused by reduction of the intramolecular disulfide bond is on average 19 kJ mol-1. Removal of the disulfide bridge by genetic substitution of C23 for valine resulted in a stable immunoglobulin domain in the context of the stabilizing Y32H amino acid exchange; again, REIv-C23V/Y32H has 18 kJ mol-1 less folding stability than REIv-Y32H. The data are consistent with the notion that all variants studied have the same overall three-dimensional structure with the disulfide bridge opened or closed. CONCLUSIONS: A comparison of the magnitude of the stabilizing effect exerted by the disulfide bond and the length of the mainchain loop framed by it suggests lowering of the entropy of the unfolded state as the sole source of the effect. Disulfide bonds are not necessary for proper folding of immunoglobulin variable domains and can be removed, provided the loss of folding stability is at least partly compensated by stabilizing amino acid exchanges.

Membrane insertion of the bacterial signal transduction protein ToxR and requirements of transcription activation studied by modular replacement of different protein substructures.

The Vibrio cholerae protein ToxR is an integral membrane protein that acts as a transcription activator in response to environmental signals; it controls expression of toxin genes ctxA and ctxB, along with a variety of other genes related to pathogenicity. Here it is shown that: (i) ToxR has a modular architecture and that activation of transcription starting at the ctx promoter depends strictly on dimerization of the periplasmic ToxR domain; (ii) the transmembrane (TM) region of ToxR is sufficient as a topogenic signal but not for stable membrane anchoring of the protein; (iii) the TM region has no special function in signal transduction and (iv) a proline residue located within the TM region minimizes background transcription activation, most plausibly by reducing TM-TM interaction. Possible applications of ToxR as a technical tool for analysing protein-protein interactions between pairs of arbitrary TM domains are discussed.

Immunoglobulin mutant library genetically screened for folding stability exploiting bacterial signal transduction.

A model repertoire of variants of immunoglobulin kappa variable domain REIv with different folding stabilities was generated by oligonucleotide-directed randomization of position 29, a key conserved residue of hypervariable loop 1. Fused to ToxR', the membrane-anchored cytoplasmic domain of the Vibrio cholerae ToxR transcription activator, different members of the library induce different levels of transcription from the ctx promoter in Escherichia coli. Differences in transcription activation correlate positively with folding stabilities of the corresponding REIv domains. Since conformationally stabilized REIv derivatives elicit a dark red colony phenotype on EMB-lactose indicator plates, this procedure constitutes a genetic screen for immunoglobulin folding stability.

Substrate preferences of Vsr DNA mismatch endonuclease and their consequences for the evolution of the Escherichia coli K-12 genome.

The substrate spectrum of Vsr DNA mismatch endonuclease of Escherichia coli K-12 was investigated using fluorescence-labelled oligonucleotide substrates and a DNA sequencer for detection and quantification of substrates and reaction products. Fourteen substrates were found to be processed by the enzyme, which differ in one or two positions from the canonical pentanucleotide sequence CTA/TGG (T mismatched to G). Relative second-order rate constants of these substrates were determined in groups of four by multiple substrate kinetics and compared to the underresentation of the corresponding pentanucleotides in the E. coli K-12 genome. The high quality of correlation further establishes active mutagenesis by VSP repair as a significant driving force of the evolution of the E. coli K-12 genome and provides clues to its possible selective value.

Two-chain bacteriorhodopsin synthesized by Schizosaccharomyces pombe.

Bacteriorhodopsin (BR), the light-driven proton pump of Halobacterium salinarium purple membrane, was produced in functional form as a two-chain protein by simultaneous expression in the fission yeast Schizosaccharomyces pombe of two separate structural genes, one coding for an aminoterminal BR fragment encompassing the first two transmembrane helices of BR, the other coding for the remainder of the protein. The fragments assemble spontaneously in vivo to yield functional BR which can be purified by immobilized metal ion affinity chromatography.

A soluble immunoglobulin variable domain without a disulfide bridge: construction, accumulation in the cytoplasm of E. coli, purification and physicochemical characterization.

Two amino acid exchanges (Y32H and C23V) were introduced sequentially into the immunoglobulin REIV, a human kappa variable domain. The first exchange stabilizes the folded state of the domain by 4.6 kJ/mol (1.1 kcal/mol), the second abolishes the central disulfide bridge and destabilizes the folded domain by 17.5 kJ/mol (4.2 kcal/mol). Introduction of the stabilizing exchange first is a necessary pre-requisite to the removal of the central disulfide bridge without collapse of the fold. The double mutant REIV-C23V/Y32H can be accumulated in the cytoplasmatic compartment of the E. coli cell, a finding that opens new possibilities in antibody engineering.

Production of the immunoglobulin variable domain REIv via a fusion protein synthesized and secreted by Staphylococcus carnosus.

REIv--the variable domain of an immunoglobulin x light chain--was produced by heterologous gene expression in a Gram-positive bacterium, purified to homogeneity and characterized. A host/vector combination based on secretion of Staphylococcus hyicus lipase by Staphylococcus carnosus was exploited. A gene encoding a fusion protein, composed of an aminoterminal portion of the pre-pro-peptide of S. hyicus lipase, a hexahistidine affinity tag, followed by the recognition sequence of IgA protease and REIv was constructed. Expression of the fusion gene in S. carnosus causes selective secretion and accumulation of a soluble fusion protein in the culture medium (5-10 mg/l), which can be purified from the supernatant by immobilized metal ion affinity chromatography (IMAC). REIv is released from the fusion protein with an additional threonine and proline residue at the aminoterminus (REIvTP) by site-specific cleavage with IgA protease and can be separated from the hexahistidine-tagged fusion partner and the protease by a second passage through an IMAC gel matrix. Like authentic REIv, the isolated protein (> 1 mg/l culture medium) migrates as a dimer in gel filtration chromatography and undergoes cooperative, reversible unfolding in urea. The isolated immunoglobulin REIvTP and authentic REIv have indistinguishable free energies of unfolding (approx. 26 kJ/mol, 6.3 kcal/mol).

Dimerization of Bence Jones proteins: linking the rate of transcription from an Escherichia coli promoter to the association constant of REIV.

Homodimers of immunoglobulin VL domains are minimal models of antibodies in that they display an ensemble of six hypervariable loops. Bence Jones protein REI is a mixture of a complete kappa light chain and the corresponding variable domain (REIV). The known three-dimensional structure of the REIV dimer (Epp et al., 1975, Biochemistry 14, 4943-4952) provides a basis for studying dimer stabilization by protein engineering. Mutant REIV-L94H was constructed and shown to have an equilibrium constant of dimerization about one order of magnitude higher than wildtype REIV. By fusing REIV and variants to the aminoterminal part of the Vibrio cholerae ToxR regulator protein (Miller et al., 1987, Cell 48, 271-279), a transcriptional signal in E. coli can be derived from REIV homodimer formation constant. The system senses dimerization of the immunoglobulin part of the fusion protein, located in the periplasmatic space, and transduces the signal as transcriptional activation to a ctx::lacZ gene construct integrated into the E. coli chromosome. There is positive correlation between the propensities of homodimer formation and the rate of transcriptional initiation at the ctx promoter. Since beta-galactosidase levels can easily be measured colorimetrically in crude cell lysates of a large number of clones using an ELISA reader, this procedure constitutes all elements required for a genetic screen in E. coli for immunoglobulin variants with altered association constants.

Properties of bacteriorhodopsin derivatives constructed by insertion of an exogenous epitope into extra-membrane loops.

Bacteriorhodopsin (BR) is folded into a bundle of seven alpha-helices which is embedded in the cellular membrane of Halobacterium salinarium; these helices are connected by short extra-membrane loops, three on the cytoplasmic side and three on the outside. Oligonucleotide-directed insertion or replacement mutagenesis was used to integrate the C-terminal sequence (13 amino acids long) of Sendai virus L-protein individually into each of the six helix-connecting loops. The altered gene products were obtained by expression of the mutant genes in either Escherichia coli or Schizosaccharomyces pombe and were used to reconstitute BR in proteoliposomes. In four cases (altered loops B/C, C/D, D/E or E/F), the mutant BRs were found to be fully functional as judged by light-driven proton pumping and photocycle kinetics. Within the four functional BR variants, recognition of the viral epitope by a monoclonal antibody is restricted to modified loops B/C and E/F. Immunogold staining of S.pombe cells producing either of the two latter BR variants shows that the protein is distributed among various cellular membranes but is not present in mitochondrial membranes. Sequence alteration of loop A/B or F/G resulted in loss of function, most plausibly due to a folding defect of the respective proteins. These results on the one hand document differences in structural importance of the various BR extra-membrane loops and on the other hand open the door to the construction of multifunctional membrane proteins via loop replacement mutagenesis of BR.

Peptide production by a combination of gene expression, chemical synthesis, and protease-catalyzed conversion.

We describe a new approach for the production of peptides using a combination of recombinant DNA technology, chemical synthesis, and proteinase-catalyzed processing. An artificial substance P-precursor is produced as a beta-galactosidase (1-459) fusion protein containing nine copies of the decapeptide sequence Arg-Leu-Arg-Arg-Pro-Lys-Pro-Gln-Gln-Phe. The fusion protein accumulates in E. coli as insoluble inclusion bodies which are easily isolated and purified. The decapeptide blocks are selectively cleaved from the insoluble fusion protein by alpha-chymotrypsin. Alternatively, a dodecapeptide ester is produced when a dipeptide ester is included in the chymotrypsin reaction mixture. This peptide ester is converted converted to substance P by papain-catalyzed acyl transfer and subsequent tryptic cleavage. These results demonstrate that peptides can be readily produced by a combination of recombinant DNA technology and proteinase-catalyzed conversion. The approach allows incorporation of groups other than natural amino acids into oligo- and polypeptides.