Naturally occurring circular permutations in proteins.

A pair of proteins is defined to be related by a circular permutation if the N-terminal region of one protein has significant sequence similarity to the C-terminal of the other and vice versa. To detect pairs of proteins that might be related by circular permutation, we implemented a procedure based on a combination of a fast screening algorithm that we had designed and manual verification of candidate pairs. The screening algorithm is a variation of a dynamic programming string matching algorithm, in which one of the sequences is doubled. This algorithm, although not guaranteed to identify all cases of circular permutation, is a good first indicator of protein pairs related by permutation events. The candidate pairs were further validated first by application of an exhaustive string matching algorithm and then by manual inspection using the dotplot visual tool. Screening the whole Swissprot database, a total of 25 independent protein pairs were identified. These cases are presented here, divided into three categories depending on the level of functional similarity of the related proteins. To validate our approach and to confirm further the small number of circularly permuted protein pairs, a systematic search for cases of circular permutation was carried out in the Pfam database of protein domains. Even with this more inclusive definition of a circular permutation, only seven additional candidates were found. None of these fitted our original definition of circular permutations. The small number of cases of circular permutation suggests that there is no mechanism of local genetic manipulation that can induce circular permutations; most examples observed seem to result from fusion of functional units.

A simple algorithm for detecting circular permutations in proteins.

MOTIVATION: Circular permutation of a protein is a genetic operation in which part of the C-terminal of the protein is moved to its N-terminal. Recently, it has been shown that proteins that undergo engineered circular permutations generally maintain their three dimensional structure and biological function. This observation raises the possibility that circular permutation has occurred in Nature during evolution. In this scenario a protein underwent circular permutation into another protein, thereafter both proteins further diverged by standard genetic operations. To study this possibility one needs an efficient algorithm that for a given pair of proteins can detect the underlying event of circular permutations. A possible formal description of the question is: given two sequences, find a circular permutation of one of them under which the edit distance between the proteins is minimal. A naive algorithm might take time proportional to N3 or even N4, which is prohibitively slow for a large-scale survey. A sophisticated algorithm that runs in asymptotic time of N2 was recently suggested, but it is not practical for a large-scale survey. RESULTS: A simple and efficient algorithm that runs in time N2 is presented. The algorithm is based on duplicating one of the two sequences, and then performing a modified version of the standard dynamic programming algorithm. While the algorithm is not guaranteed to find the optimal results, we present data that indicate that in practice the algorithm performs very well. AVAILABILITY: A Fortran program that calculates the optimal edit distance under circular permutation is available upon request from the authors. CONTACT: ron@biocom1.ls.biu.ac.il.

Computational analysis of the amino acid residue sequences of amaranth and some other proteins.

Amaranth belongs to a nutritious class of pseudocereals. The well balanced amino acid composition of amaranth compared with those of major cereals would indicate that it deserves a quantitative study of its chemical properties. This work was undertaken to compare Amaranthus (A.) caudatus with a number of other plants on the basis of the sequences of various proteins and the composition of their alcohol-soluble protein mixture and glutelins. Alcohol-soluble proteins were extracted with 55% isopropanol (2-ProOH) + 5% 2-mercaptoethanol (2-ME) and glutelin fractions were obtained with borate buffer + 3% 2-ME + 0.5% sodium dodecyl sulfate (SDS), pH 10. Protein fractions were then electrophoreded on sodium dodecyl sulfate polyacrylamide gels (SDS-PAGE). FASTA and TFASTA programmes were used for comparison of amino acid sequences. Dot matrix analysis and secondary structure predictions which were drawn by Plot-structure, were taken from the GCG package. Electrophoretic tests failed to indicate significant correlation between prolamins from cereals and other plants with the alcohol-soluble fractions from amaranth, proving that these proteins cannot represent the main fraction in amaranth. On the other hand, glutelins shared some common electrophoretic bands with other cereals and showed some identity by SDS-PAGE. Amino acid sequences of A. caudatus (100% identity) had degrees of similarity in the range of 71.4 to 52.2% with rice, garden pea, jobs' tears, maize, and yam. Rice glutelin had similarity in the range of 93.3% to 44.8% with oats, soybean, and pea. Secondary structures of A. caudatus (using conservative amino acid replacement), jobs' tears and rice glutelins, oat globulin, and pea legumin sequences were predicted. Some relationship was shown among electrophoretic patterns of alcohol-soluble proteins and glutelins of A. caudatus.

Extending crystallographic information with semiempirical quantum mechanics and molecular mechanics: a case of aspartic proteinases.

The results of crystallographic analysis of a complex between an aspartic proteinase, endothiapepsin, and an inhibitor have been extended through the assignment of protons in the active site, to study various steps in the reaction with a substrate. Mechanistic implications are suggested as a consequence of semiempirical quantum mechanical calculations, indicating that most of the activation energy is required to bring the substrate from an initial binding mode to close distance to a water molecule.

Genetic engineering, isolation and characterization of a truncated Escherichia coli elongation factor Tu comprising domains 2 and 3.

A deletion mutant of a plasmid born Escherichia coli tufA gene, which codes for a truncated elongation factor Tu comprising domains 2 and 3, has been constructed by genetic engineering. This gene was overexpressed in E. coli, and a polypeptide representing the truncated elongation factor Tu was isolated, purified to near homogeneity, crystallized and characterized physico-chemically as well as biochemically. Circular dichroism spectroscopy and limited tryptic digestion demonstrate that the isolated domain pair 2 and 3 behaves like an independent folding unit which adopts a similar secondary and most likely, tertiary, structure to that present in the intact elongation factor Tu. However, the isolated domain pair 2 and 3 does not interact with aminoacyl-tRNA or the antibiotic kirromycin, two ligands which were shown previously by cross-linking experiments to be in contact with amino acid residues located in domains 1 and 2, and domain 3, respectively. The results suggest that the isolated domain pair 2 and 3 by itself forms too few contacts with these ligands to form a stable complex. Furthermore, the data suggest that domain 1 in intact EF-Tu, in a subtle but nevertheless decisive manner, alters the conformation of the other two domains in such a way that all three domains cooperatively create a high affinity binding site for aminoacyl-tRNA and the antibiotic kirromycin.

Accuracy of the EcoRI restriction endonuclease: binding and cleavage studies with oligodeoxynucleotide substrates containing degenerate recognition sequences.

We have synthesized a series of 18 nonpalindromic oligodeoxynucleotides that carry all possible base changes within the recognition sequence of EcoRI. These single strands can be combined with their complementary single strands to obtain all possible EcoRI sequences (left), or they can be combined with a single strand containing the canonical sequence to obtain double strands with all possible mismatches within the recognition sequence (right): (sequence; see text) The rate of phosphodiester bond cleavage of these oligodeoxynucleotides by EcoRI was determined in single-turnover experiments under normal buffer conditions in order to find out to what extent the canonical recognition site can be distorted and yet serve as a substrate for EcoRI. Our results show that oligodeoxynucleotides containing mismatch base pairs are in general more readily attacked by EcoRI than oligodeoxynucleotides containing EcoRI sites and that the rates of cleavage of the two complementary strands of degenerate oligodeoxynucleotides are quite different. We have also determined the affinities of these oligodeoxynucleotides to EcoRI. They are higher for oligodeoxynucleotides carrying a mismatch within the EcoRI recognition site than for oligodeoxynucleotides containing an EcoRI site but otherwise do not correlate with the rate with which these oligodeoxynucleotides are cleaved by EcoRI. Our results allow details to be given for the probability of EcoRI making mistakes in cleaving DNA not only in its recognition sequence but also in sequences closely related to it. Due to the fact that the rates of cleavage in the two strands of a degenerate sequence generally are widely different, these mistakes are most likely not occurring in vivo, since nicked intermediates can be repaired by DNA ligase.

Probing the function of individual amino acid residues in the DNA binding site of the EcoRI restriction endonuclease by analysing the toxicity of genetically engineered mutants.

We have developed an assay that allows analysis of the activity of EcoRI restriction endonuclease (ENase) and its mutants in vivo. This assay is based on the fact that wild type (wt) EcoRI ENase is toxic for Escherichia coli cells not expressing the EcoRI methyltransferase (MTase). The viability factor defined by the ratio of the viable counts of E. coli cultures having or not having expressed the ecoRIR gene for a defined time is 10(-6) for wt EcoRI ENase and close to one for a totally inactive EcoRI ENase mutant. While the EcoRI MTase (M.EcoRI) provides substantial protection against the toxic effects of the wt EcoRI ENase and several of the mutants, some mutants become more toxic in the presence of M.EcoRI. Twenty-four different DNA-binding-site mutants of EcoRI ENase were characterized in their activity in vivo with this assay. The results obtained allow us to conclude that the structural integrity of the region at and around aa 200 seems to be very critical for the enzymatic function of EcoRI ENase: nonconservative replacements there lead to viability factors of 1-10(-2). While our results indicate that the region around aa 144 and 145 is also involved in the EcoRI ENase-catalyzed reaction, it is also evident that the effects of mutation there are not as large: viability factors of approx. 10(-3) are obtained even for drastic replacements. These results are discussed in the light of the x-ray structure analysis of an EcoRI ENase-DNA recognition complex.

Fluorescence stopped-flow kinetics of the cleavage of synthetic oligodeoxynucleotides by the EcoRI restriction endonuclease.

We have investigated in fluorescence stopped-flow and temperature-jump experiments the EcoRI-catalyzed cleavage of synthetic palindromic tridecadeoxynucleotides which contain the EcoRI site but differ in the flanking sequences. The overall reaction can be resolved in several reactions which were analyzed by a nonlinear least-squares fitting procedure on the experimental data. The result of this analysis is a minimal scheme that describes the overall reaction in terms of the rate constants of the individual reactions. According to this scheme EcoRI and the tridecadeoxynucleotide substrates associate in the presence of Mg2+ in a nearly diffusion-controlled process. This is followed by a reaction which is or includes the cleavage of the first phosphodiester bond. There is no indication for a time-resolved conformational transition prior to catalysis. After cleavage of the first strand, dissociation of the nicked double strand can occur, which then rearranges to the original palindromic double-stranded substrate and is bound again by the enzyme. Alternatively, the nicked double strand can be cleaved in the second strand. This reaction is followed by product release from the enzyme. The magnitude of the individual rate constants depends on the substrate used; the differences explain the preference of EcoRI for substrates that contain AT as compared to GC base pairs next to the recognition site.

Changing the hydrogen-bonding potential in the DNA binding site of EcoRI by site-directed mutagenesis drastically reduces the enzymatic activity, not, however, the preference of this restriction endonuclease for cleavage within the site-GAATTC-.

According to the X-ray structure analysis of an EcoRI-oligodeoxynucleotide complex [McClarin et al. (1986) Science 234, 1526], sequence specificity is mediated by 12 hydrogen bonds, 6 from each of the two identical subunits of the dimeric enzyme to the recognition site -GAATTC-: Arg200 forms two hydrogen bonds with guanine, while Glu144 and Arg145 form four hydrogen bonds to adjacent adenine residues. Changing the hydrogen-bonding potential at the recognition site without perturbing the rest of the interface should lead to the recognition of degenerate sequences [Rosenberg et al. (1987) in Protein Engineering (Oxender, D. L., & Fox, C. F., Eds.) pp 237-250, Liss, New York]. We have shown previously that replacing Glu144 by Gln and Arg145 by Lys affects the activity of the enzyme, not, however, its specificity [Wolfes et al. (1986) Nucleic Acids Res. 14, 9063]. We show now that also the mutation of Arg200 to Lys, the double mutation Glu144Arg145 to GlnLys, and the triple mutation Glu144Arg145Arg200 to GlnLysLys do not lead to a detectable degeneracy of the specificity of cleavage by EcoRI but significantly impair the catalytic activity of this enzyme. A detailed analysis of the steady-state kinetics of cleavage of pUC8 DNA and a tridecadeoxynucleotide substrate demonstrates that the reduction in activity for all DNA binding site mutants investigated so far is mainly due to a decrease in kcat, with the exception of the Arg200 to Lys mutant, which is only impaired in its KM.(ABSTRACT TRUNCATED AT 250 WORDS)

Genetic engineering of EcoRI mutants with altered amino acid residues in the DNA binding site: physicochemical investigations give evidence for an altered monomer/dimer equilibrium for the Gln144Lys145 and Gln144Lys145Lys200 mutants.

We have genetically engineered the Arg200----Lys mutant, the Glu144Arg145----GlnLys double mutant, and the Glu144Arg145Arg200----GlnLysLys triple mutant of the EcoRI endonuclease in extension of previously published work on site-directed mutagenesis of the EcoRI endonuclease in which Glu144 had been exchanged for Gln and Arg145 for Lys [Wolfes et al. (1986) Nucleic Acids Res. 14, 9063]. All these mutants carry modifications in the DNA binding site. Mutant EcoRI proteins were purified to homogeneity and characterized by physicochemical techniques. All mutants have a very similar secondary structure composition. However, whereas the Lys200 mutant is not impaired in its capacity to form a dimer, the Gln144Lys145 and Gln144Lys145Lys200 mutants have a very much decreased propensity to form a dimer or tetramer depending on concentration as shown by gel filtration and analytical ultracentrifugation. This finding may explain the results of isoelectric focusing experiments which show that these two mutants have a considerably more basic pI than expected for a protein in which an acidic amino acid was replaced by a neutral one. Furthermore, while wild-type EcoRI and the Lys200 mutant are denatured in an irreversible manner upon heating to 60 degrees C, the thermal denaturation process as shown by circular dichroism spectroscopy is fully reversible with the Gln144Lys145 double mutant and the Gln144Lys145Lys200 triple mutant. All EcoRI endonuclease mutants described here have a residual enzymatic activity with wild-type specificity, since Escherichia coli cells overexpressing the mutant proteins can only survive in the presence of EcoRI methylase. The detailed analysis of the enzymatic activity and specificity of the purified mutant proteins is the subject of the accompanying paper [Alves et al. (1989) Biochemistry (following paper in this issue)].

Modulation of the affinity of the single-stranded DNA-binding protein of Escherichia coli (E. coli SSB) to poly(dT) by site-directed mutagenesis.

A vector for site-directed mutagenesis and overproduction of the Escherichia coli single-stranded-DNA-binding protein (E. coli SSB) was constructed. An E. coli strain carrying this vector produces up to 400 mg pure protein from 25 g wet cells. The vector was used to mutate specifically the Phe60 residue of E. coli SSB. Phe60 had been proposed to be located near the single-stranded-DNA-binding site. Substitution of the Phe60 residue by Val, Ser, Leu, His, Tyr and Trp gave proteins with no or only minor conformational changes, as detected by NMR spectroscopy. The affinity of the mutant E. coli SSB proteins for single-stranded DNA decreased in the order Trp greater than Phe (wild-type) greater than Tyr greater than Leu greater than His greater than Val greater than Ser, leading to the conclusion that position 60 is a site of hydrophobic interaction of the protein with DNA.

Analysis of the recognition mechanism involved in the EcoRV catalyzed cleavage of DNA using modified oligodeoxynucleotides.

We have prepared a series of undecadeoxynucleotides that contain changes in the functional group pattern present within the EcoRV recognition site - GATATC-. Oligonucleotides were synthesized on solid phase using normal and modified beta-cyanoethylphosphoramidites and analyzed in steady state cleavage experiments with the EcoRV restriction endonuclease. The following groups appear to interact strongly with the enzyme, since their modification or substitution renders the oligonucleotides refractory to cleavage: the exocyclic NH2-groups of both A residues, the N7 of the first A residue, the exocyclic NH2-group of the C residue and the CH3-groups of both T residues. The exocyclic NH-group of the G residue supports effective recognition, since its absence lowers the kcat of the cleavage reaction. The N7 of the second A residue and the C5 position of the C residue apparently are not recognized by EcoRV; their substitution by -CH- or modification with -Br or -CH3, resp., does not considerably change the rate of cleavage. All oligonucleotides investigated compete with the unmodified substrate for binding to the enzyme. We conclude that EcoRV recognizes its substrate presumably through hydrogen bonds to the exocyclic NH2-group and the N7 of the first A residue, the exocyclic NH2-groups of the second A and the C residue, as well as through hydrophobic interactions with both T residues.

Site directed mutagenesis experiments suggest that Glu 111, Glu 144 and Arg 145 are essential for endonucleolytic activity of EcoRI.

We have constructed a plasmid (pRIF 309+) carrying the EcoRI restriction endonuclease gene and the f1 origin of replication. Upon transformation of this plasmid into E. coli and infection with bacteriophage f1 single stranded plasmids are produced which can be used for sequencing and site directed mutagenesis. Using this single stranded DNA and synthetic oligodeoxynucleotides we have introduced point mutations at defined positions of the EcoRI gene. Since in pRIF309+ the EcoRI gene is under the control of the pL-promoter, high level expression of the mutated EcoRI gene could be obtained upon induction. Mutant EcoRI enzymes were purified to homogeneity and characterized in structural and functional terms. Our results demonstrate that the Glu 111----Gln, Glu 144----Gln and Arg 145----Lys -mutants adopt a very similar conformation as the wild type enzyme, but have by two orders of magnitude smaller specific activities than the wild type enzyme, mainly due to a reduction of the Vmax-value.

Cross-linking of bromodeoxyuridine-substituted oligonucleotides to the EcoRI and EcoRV restriction endonucleases.

We have synthesized several self-complementary oligodeoxynucleotides which contain bromodeoxyuridine in various positions within and outside of the recognition sequence for the EcoRI and EcoRV restriction endonucleases. These oligodeoxynucleotides are cleaved in the presence of Mg2+ by their respective enzyme. Upon irradiation by long-wavelength ultraviolet light and in the absence of Mg2+ they are cross-linked in low yield to their enzymes, forming 1:1 and 1:2 (oligodeoxynucleotide:enzyme subunit) adducts. Cross-linking occurs with both specific and non-specific complexes. With EcoRI the site of cross-linking was determined to be at or close to Met-137, i.e. in a region of the molecule implicated by other studies from our laboratory [Scholtissek et al. (1986) J. Biol. Chem. 261, 2228-2234] in the binding and cleavage of the substrate.

Role of thymidine residues in DNA recognition by the EcoRI and EcoRV restriction endonucleases.

We have synthesized a series of oligonucleotides containing the EcoRI (GAATTC) or EcoRV (GATATC) recognition site within which or adjacent to which thymidine was substituted by uridine or derivatives of uridine. The effects of these substitutions on the rate of the EcoRI and EcoRV catalyzed cleavage reaction were investigated. Our results show that most of the substitutions within the site are quite well tolerated by EcoRI, not, however, by EcoRV. We conclude that the thymin residues most likely are not directly involved in the recognition process of the EcoRI reaction. In contrast, they are major points of contact, between substrate and enzyme in the EcoRV reaction. The effects of substitutions in the position adjacent to the recognition site is also markedly different for EcoRI and EcoRV. Here, EcoRI seems to be considerably more selective than EcoRV.

A comparison of the structural requirements for DNA cleavage by the isoschizomers HaeIII, BspRI and BsuRI.

We have investigated the structural requirements for DNA cleavage by the isoschizomers HaeIII, BspRI and BsuRI which recognize the sequence -d(GGCC)-. For this purpose decadeoxynucleotides were synthesized by the solid-phase phosphotriester method and purified by high-performance liquid chromatography. The kinetics of cleavage of these oligodeoxynucleotides were determined for the three isoschizomers with the following results. The sequence adjacent to the recognition site strongly influences the rate of cleavage. The preference is qualitatively the same for all three enzymes: AGGCCT greater than TGGCCA greater than GGGCCC approximately equal to CGGCCG, and follows the thermal stability of the different decanucleotides. Substitutions within the recognition site, namely dI for dG and dU for dC, affect the rate of cleavage differently for the three enzymes. The results can be rationalized in terms of an interaction of HaeIII with the major and minor groove of the DNA, of BspRI mainly with the minor groove and of BsuRI with the major groove of DNA. It is obvious from our data that the mechanism of recognition of the same site is different for the three isoschizomers.