Distribution of potential type II restriction sites (palindromes) in prokaryotes

Restriction-modification systems are used as a defensive mechanism against inappropriate invasion of foreign DNA. The recognition sequences for the common type II restriction enzymes and their corresponding methylases are usually palindromes. In this study, we identified the most over- and underrepresented words in DNA of four bacteria: Escherichia coli, Bacillus subtilis, Clostridium perfringens, and Pseudomonas aeruginosa. Using maximum order Markov chain analysis, we found that palindromic words were most often more underrepresented than their non-palindromic counterparts. No strict rule for the intragenic palindrome content could be derived, but for three of the bacteria there was a weak correlation between codon usage bias and palindrome content. A clear drop in palindrome counts was observed in the Shine-Dalgarno region for B. subtilis and C. perfringens, but not in E. coli or P. aeruginosa. It was also shown that palindromes in eubacteria and archaebacteria seem to occur slightly more infrequently than expected on the basis of the genomic GC-content, but some exceptions to this principle exist.     

Analysis of distal flanking regions of maize 19-kDa zein genes

Two genomic fragments from maize, each containing a 19-kDa zein gene with extensive flanking regions, have been sequenced and examined by computer-aided analysis and Southern blotting techniques. Sequence analysis of the distal flanking sequences has revealed interesting sequence motifs, some not seen before. In particular, four nearly identical, G + C-rich, 17 to 21-bp perfect palindromes were found clustered in a 133-bp stretch lying 2 kb upstream from the zein-coding region in the genomic clone pMS2. These palindromic sequences exhibit other interesting features, including a precise spatial organization with respect to each other, and their proximity to several other repeated motifs in the same region. Southern blot analysis indicates that these palindromes, or closely related sequences, are found frequently in the maize genome. Possible secondary structures for the palindrome units are presented, which resemble functionally important sequences found upstream from other eukaryotic genes.     

Molecular cloning and expression of Bacillus licheniformis beta-lactamase gene in Escherichia coli and Bacillus subtilis.

The chromosomal beta-lactamase (penicillinase, penP) gene from Bacillus licheniformis 749/C has been cloned in Escherichia coli. The locations of the target sites for various restriction enzymes on the 4.2-kilobase EcoRI fragment were determined. By matching the restriction mapping data with the potential nucleotide sequences of the penP gene deduced from known protein sequence, we established the exact position of the penP gene on the fragment. A bifunctional plasmid vector carrying the penP gene, plasmid pOG2165, was constructed which directs the synthesis of the heterologous beta-lactamase in both E. coli and Bacillus subtilis hosts. The protein synthesized in E. coli and B. subtilis is similar in size to the processed beta-lactamase made in B. licheniformis. Furthermore, the beta-lactamase made in B. subtilis is efficiently secreted by the host into the culture medium, indicating that B. subtilis is capable of carrying out the post-translational proteolytic cleavage(s) to convert the membrane-bound precursor enzyme into the soluble extracellular form.     

Underrepresentation of short palindromes in Selenomonas ruminantium DNA: evidence for horizontal gene transfer of restriction and modification systems?

Molecular analysis of isolates of the rumen bacterium Selenomonas ruminantium revealed a high variety and frequency of site-specific (restriction) endonucleases. While all known S. ruminantium restriction and modification systems recognize hexanucleotide sequences only, consistently low counts of both 6-bp and 4-bp palindromes were found in DNA sequences of S. ruminantium. Statistical analysis indicated that there is some correlation between the degree of underrepresentation of tetranucleotide words and the number of known restriction endonucleases for a given sequence. Control analysis showed the same correlation in lambda DNA but not in human adenovirus DNA. Based on the data presented, it could be proposed that there is a much higher historical occurrence of restriction and modification systems in S. ruminantium and (or) frequent horizontal gene transfer of restriction and modification gene complexes.     

A new cloning system for Bacillus subtilis comprising elements of phage, plasmid and transposon vectors

A new cloning system for Bacillus subtilis was devised which makes use of a combination of Tn917-containing phage SP beta derivatives and Tn917-containing Escherichia coli-B. subtilis shuttle plasmids. This system allows the initial cloning of genes in single copy, via 'prophage transformation', with a selection for complementation of mutational defects in B. subtilis hosts and permits subsequent transfer of the cloned material by homologous recombination to low-copy and high-copy vectors that replicate in both B. subtilis and E. coli. Because cloned sequences are adjacent to pB322-derived DNA in the recombinant phages, inserts can also be 'rescued' directly from the phage DNA after digestion with appropriate restriction enzymes, circularization of the fragments by ligation and transformation of an E. coli recipient. Two genomic libraries of B. subtilis chromosomal Sau3A-generated partial-digest fragments in the size ranges of 5-8 kb and 8-10 kb were constructed and screened for the complementation of mutations aroI906, cysA14, dal-1, glyB133, metC3, purA16, purB33, thrA5, trpC2 and recE4. In all cases, specialized transducing phages carrying inserts that complemented the selected markers were recovered. Inserts complementing the dal-1 and trpC2 mutations could be transferred from recombinant phages to Tn917-containing plasmids by homologous recombination without in vitro subcloning. Another insert complementing the purB33 mutation was rescued directly into E. coli from a recombinant phage DNA.     

Palindromes in proteins

Palindromes in DNA consist of nucleotides sequences that read the same from the 5'-end to the 3'-end, and its double helix is related by twofold axis. They occur in genomes of all organisms and have various functions. For example, restriction enzymes often recognize palindromic sequences of DNA. Palindromes in telomeres are crucial for initiation of replication. One can ask the questions, Do palindromes occur in protein, and if so, what function they play? We have searched the protein SWISSPROT database for palindromic sequences. A great number (26%) of different protein palindromes were found. One example of such protein is systemin, an 18-amino-acid-long peptide. It contains palindrome in its beta-sheet domain that interacts with palindromic fragment of DNA. The other palindrome containing protein is cellular human tumor suppressor p53. Oligonucleotide LTI-ITL has been observed in the crystal structure and is located close to a DNA recognizing domain. As the number of possible palindromic sequences of a given length is far much greater for proteins (20N) than for nucleic acids (4N), the study on their role seems to be an exciting challenge. Our results have clearly showed that palindromes are frequently occurring motives in proteins. Moreover, even very few examples that we have examined so far indicate the importance of further studies on protein palindromes.     

Palindromes in Proteins

Palindromes in DNA consist of nucleotides sequences that read the same from the 5′-end to the 3′-end, and its double helix is related by twofold axis. They occur in genomes of all organisms and have various functions. For example, restriction enzymes often recognize palindromic sequences of DNA. Palindromes in telomeres are crucial for initiation of replication. One can ask the questions, Do palindromes occur in protein, and if so, what function they play? We have searched the protein SWISSPROT database for palindromic sequences. A great number (26%) of different protein palindromes were found. One example of such protein is systemin, an 18-amino-acid-long peptide. It contains palindrome in its β-sheet domain that interacts with palindromic fragment of DNA. The other palindrome containing protein is cellular human tumor suppressor p53. Oligonucleotide LTIITL has been observed in the crystal structure and is located close to a DNA recognizing domain. As the number of possible palindromic sequences of a given length is far much greater for proteins (20^N) than for nucleic acids (4^N), the study on their role seems to be an exciting challenge. Our results have clearly showed that palindromes are frequently occurring motives in proteins. Moreover, even very few examples that we have examined so far indicate the importance of further studies on protein palindromes.     

Molecular evolution of bacteriophages: evidence of selection against the recognition sites of host restriction enzymes

Restriction enzymes produced by bacteria serve as a defense against invading bacteriophages, and so phages without other protection would be expected to undergo selection to eliminate recognition sites for these enzymes from their genomes. The observed frequencies of all restriction sites in the genomes of all completely sequenced DNA phages (T7, lambda, phi X174, G4, M13, f1, fd, and IKe) have been compared to expected frequencies derived from trinucleotide frequencies. Attention was focused on 6-base palindromes since they comprise the typical recognition sites for type II restriction enzymes. All of these coliphages, with the exception of lambda and G4, exhibit significant avoidance of the particular sequences that are enterobacterial restriction sites. As expected, the sequenced fraction of the genome of phi 29, a Bacillus subtilis phage, lacks Bacillus restriction sites. By contrast, the RNA phage MS2, several viruses that infect eukaryotes (EBV, adenovirus, papilloma, and SV40), and three mitochondrial genomes (human, mouse, and cow) were found not to lack restriction sites. Because the particular palindromes avoided correspond closely with the recognition sites for host enzymes and because other viruses and small genomes do not show this avoidance, it is concluded that the effect indeed results from natural selection.      

Molecular cloning of the terminal hairpin of vaccinia virus DNA as an imperfect palindrome in an Escherichia coli plasmid

The genome of vaccinia virus is a linear duplex molecule of approximately 185 kb with hairpins at each end that link the complementary strands. The hairpins, which exist in two forms that are inverted and complementary in sequence, were isolated as XbaI restriction fragments and converted to a linear intermolecular duplex structure by denaturation and reannealing. The latter was then stably cloned as a 142-bp imperfect palindrome in an Escherichia coli plasmid. The insert was excised from the plasmid and the palindrome was extended on both sides by ligating it to the adjacent vaccinia virus DNA segment. The resulting fragment was cloned as a 278-bp imperfect palindrome. Restriction endonuclease analysis and DNA sequencing indicated the absence of any deletions or rearrangements. After excision from the plasmid, the palindrome was converted by heating and rapid cooling to the original two hairpin forms. In this manner, large quantities of vaccinia virus telomeres may be obtained for physical and biochemical studies.     

Effects of palindrome size and sequence on genetic stability in the bacteriophage phi X174 genome

By inserting palindromes of varying length and sequence into a non-essential region of the bacteriophage phi X174 genome we have investigated the effect of palindrome size and sequence on their genetic stability. Multimers of increasing size of the EcoRI linker CCGAATTCGG (E), the BamHI linker CCGGATCCGG (B) or mixtures of both (E, B) were inserted into the PvuII site of a previously constructed bacteriophage strain phi X174 J-F ins6. The largest inserts that could be maintained in the genome without significant loss of genetic stability were 2B, 4E, and 4(E, B), respectively. Polymers exceeding this size could be inserted but resulted in rapid and precise deletion from the phage genome, whereby nB was more unstable than nE, and nE was more unstable than n(E, B). Analysis of the resulting deletion mutants provided evidence for two different types of deletions. The more frequent deletion arose from either type palindrome and removed nucleotides in blocks of ten base-pairs (one linker unit), but only from the palindromic sequence, and always left at least an 18 base-pair long palindrome (one linker plus 8 neighboring base-pairs) behind. The less frequently occurring deletions arose only from nB type palindromes, removing the complete palindromic sequence plus adjacent nucleotides. At least the first type of deletion occurred in the absence of recA activity. Our results show a correlation between the sequence, as well as size, and the genetic stability of palindromes, i.e. sequences that could decrease the stability of a cruciform increased their genetic stability. This supports the theory that palindrome deletion occurs via extrusion of the palindrome into a cruciform or cruciform-like structure.     

枯草杆菌基因组中的回文结构

统计了枯草杆菌全序列中中间间隔S从0到29、侧翼序列长度L从4开始的所有回文结构,以及这些回文结构在编码区和非编码区的分布。通过分析不同S、L的回文结构的频数以及AT含量,发现枯草杆菌基因组中长的回文结构是过表达的、AT含量高并且对非编码区有偏好。     

Recombination efficiency is a quadratic function of the length of homology during plasmid transformation of Bacillus subtilis protoplasts and Escherichia coli competent cells.

Recombination efficiency between transforming and resident plasmids was 100 times higher in Bacillus subtilis protoplasts than in Escherichia coli competent cells. In both systems it varied with the square of the length of the sequences common to two plasmids within the studied range 0.96-8.8 kb in B. subtilis, 0.72-3.9 kb in E. coli. The observed exponential dependence may be characteristic of recombination occurring within a relatively short interval bordered by relatively long heterologous regions.     

Determination of bias in the relative abundance of oligonucleotides in DNA sequences.

Different statistical measures of bias of oligonucleotide sequences in DNA sequences were compared, both by theoretical analysis and according to their abilities to predict the relative abundances of oligonucleotides in the genome of Escherichia coli. The expected frequency of an oligonucleotide calculated from a maximal order Markov model was shown to be a degenerate case of the expected frequency calculated from biases of all subwords arising when noncontiguous subwords exhibit no bias. Since (at least in E. coli) noncontiguous sequences exhibit significant bias, the total compositional bias approach is expected to represent biases in genomic sequences more faithfully than Markov approaches. In fact, the efficacy of statistics based on Markov analysis even at the highest order were inferior in predicting actual frequencies of oligonucleotides to methods that factored out biases of internal subwords with gaps. Using total compositional bias as a measure of relative abundance, tetranucleotide and hexanucleotide palindromes were found to be distributed differently from nonpalindromic sequences, with their means shifted somewhat towards underrepresentation. A subpopulation of palindromic hexanucleotides, however, was highly underrepresented, and this group consisted almost entirely of targets for Type II restriction enzymes found within strains of E. coli. Sites recognized by Type I endonucleases from related strains were not markedly biased, and with pentanucleotides, palindromic and nonpalindromic sequences had nearly identical distributions. The loss of restriction sites may be explained by the free transfer of plasmids encoding restriction enzymes and episodic selection for the presence of the enzymes.     

Comparison of physical and genetic properties of palindromic DNA sequences.

Some viable palindromic DNA sequences were found to cause an increase in the recovery of genetic recombinants. Although these palindromes contained no Chi sites, their presence in cis caused apparent recA+-dependent recombination to increase severalfold. This biological property did not correlate with the physical properties of the palindromes' extrusion of cruciform structures in vitro. Thus, two unrelated palindromes with similar effects on recombination in both Escherichia coli and Pseudomonas syringae displayed quite different kinetics of cruciform formation. In plasmids of native superhelical density, one palindrome underwent rapid cruciform formation at 55 degrees C, whereas the other did not form detectable cruciforms at any temperature. A shorter palindrome with similarly rapid kinetics of cruciform formation did not affect recombination detectably. The lack of a clear relationship between physical and genetic properties was also demonstrated in the case of longer, inviable palindromes. Here we found that the degree of asymmetry required in vivo to rescue a long palindrome from inviability far exceeded that required to kinetically prohibit cruciform extrusion in vitro.     

GAP-Seq: a method for identification of DNA palindromes

Background

Closely spaced long inverted repeats, also known as DNA palindromes, can undergo intrastrand annealing to form DNA hairpins. The ability to form these hairpins results in genome instability, difficulties in maintaining clones in Escherichia coli and major problems for most DNA sequencing approaches. Because of their role in genomic instability and gene amplification in some human cancers, it is important to develop systematic approaches to detect and characterize DNA palindromes.

Results

We developed a new protocol to identify palindromes that couples the S1 nuclease treated Cot0 DNA (GAPF) with high-throughput sequencing (GAP-Seq). Unlike earlier protocols, it does not involve restriction enzymatic digestion prior to DNA snap-back thereby preserving longer DNA sequences. It also indicates the location of the novel junction, which can then be recovered. Using MCF-7 breast cancer cell line as the proof-of-principle analysis, we have identified 35 palindrome candidates and physically characterized the top 5 candidates and their junctions. Because this protocol eliminates many of the false positives that plague earlier techniques, we have improved palindrome identification.

Conclusions

The GAP-Seq approach underscores the importance of developing new tools for identifying and characterizing palindromes, and provides a new strategy to systematically assess palindromes in genomes. It will be useful for studying human cancers and other diseases associated with palindromes.

Electronic supplementary material

The online version of this article (doi:10.1186/1471-2164-15-394) contains supplementary material, which is available to authorized users.     

Cloning and characterization of biotin biosynthetic genes of Kurthia sp

The biotin biosynthesis genes of Kurthia sp., which is an aerobic gram-positive bacterium, were cloned from Kurthia sp. 538-KA26 and characterized. Eleven biotin biosynthetic genes have been identified in Kurthia sp. Kurthia sp. has two genes coding for KAPA synthase, bioF and bioFII, and also has two genes coding for BioH protein, bioH and bioHII. In addition, three genes, orf1, orf2, and orf3, whose functions are unknown, were found in the biotin gene clusters of Kurthia sp. The bioA, bioD, and orf1 genes are arranged in a gene cluster in the order orf1bioDA, and the bioB, bioF, and orf2 genes are arranged in a gene cluster in the order orf2bioFB. These gene clusters proceed to both directions; the face to face promoters and two 40-bp of palindrome sequences exist upstream of the orf1 and orf2 genes. The bioC, bioFII, and bioHII genes are arranged in a gene cluster in the order bioFIIHIIC; a 40-bp of palindrome sequence exists upstream of the bioFII gene. The bioH and orf3 genes are arranged in a gene cluster in the order bioHorf3; a palindrome sequence was not found upstream of the bioH gene. These palindrome sequences are extremely similar to each other, suggesting that the orf1bioDA, orf2bioFB, and bioFIIHIIC gene clusters are regulated by biotin. Kurthia sp. does not have the bioW gene coding pimeloyl-CoA synthase, suggesting that pimeloyl-CoA may be produced by a different pathway than that of gram-positive bacterium B. subtilis or B. sphaericus, further suggesting a modified fatty acid synthesis pathway via acetyl-CoA instead as E. coli has.     

Large DNA palindromes as a common form of structural chromosome aberrations in human cancers

Breakage-fusion-bridge cycles contribute to chromosome aberrations and generate large DNA palindromes that facilitate oncogene amplification in cancer cells. At the molecular level, large DNA palindrome formation is initiated by chromosome breaks, and genomic architecture such as short inverted repeat sequences facilitates this process in mammalian cells. However, the prevalence of DNA palindromes in cancer cells is currently unknown. To determine the prevalence of DNA palindromes in human cancer cells, we have developed a new microarray-based approach called Genome-wide Analysis of Palindrome Formation (GAPF, Tanaka et al., Nat Genet 2005; 37: 320-7). This approach is based on a relatively simple and efficient method to purify "snap-back DNA" from large DNA palindromes by intramolecular base-pairing, followed by elimination of single-stranded DNA by nuclease S1. Comparison of Genome-wide Analysis of Palindrome Formation profiles between cancer and normal cells using microarray can identify genome-wide distributions of somatic palindromes. Using a human cDNA microarray, we have shown that DNA palindromes occur frequently in human cancer cell lines and primary medulloblastomas. Significant overlap of the loci containing DNA palindromes between Colo320DM and MCF7 cancer cell lines suggests regions in the genome susceptible to chromosome breaks and palindrome formation. A subset of loci containing palindromes is associated with gene amplification in Colo320DM, indicating that the location of palindromes in the cancer genome serves as a structural platform that supports subsequent gene amplification.     

Identification of medicinal mushroom species based on nuclear large subunit rDNA sequences

The purpose of this study was to develop molecular identification method for medical mushrooms and their preparations based on the nucleotide sequences of nuclear large subunit (LSU) rDNA. Four specimens were collected of each of the three representative medicinal mushrooms used in Korea: Ganoderma lucidum, Coriolus versicolor, and Fomes fomentarius. Fungal material used in these experiments included two different mycelial cultures and two different fruiting bodies from wild or cultivated mushrooms. The genomic DNA of mushrooms were extracted and 3 nuclear LSU rDNA fragments were amplified: set 1 for the 1.1-kb DNA fragment in the upstream region, set 2 for the 1.2-kb fragment in the middle, and set 3 for the 1.3-kb fragment downstream. The amplified gene products of nuclear large subunit rDNA from 3 different mushrooms were cloned into E. coli vector and subjected to nucleotide sequence determination. The sequence thus determined revealed that the gene sequences of the same medicinal mushroom species were more than 99.48% homologous, and the consensus sequences of 3 different medicinal mushrooms were more than 97.80% homologous. Restriction analysis revealed no useful restriction sites for 6-bp recognition enzymes for distinguishing the 3 sequences from one another, but some distinctive restriction patterns were recognized by the 4-bp recognition enzymes AccII and HhaI. This analysis was also confirmed by PCR-RFLP experiments on medicinal mushrooms.     

PY54, a linear plasmid prophage of Yersinia enterocolitica with covalently closed ends

PY54 is a temperate phage isolated from Yersinia enterocolitica. Lysogenic Yersinia strains harbour the PY54 prophage as a plasmid (pY54). The plasmid has the same size (46 kb) as the PY54 genome isolated from phage particles. By electron microscopy, restriction analysis and DNA sequencing, it was demonstrated that the phage and the plasmid DNAs are linear, circularly permuted molecules. Unusually for phages of Gram-negative bacteria, the phage genome has 3'-protruding ends. The linear plasmid pY54 has covalently closed ends forming telomere-like hairpins. The equivalent DNA sequence of the phage genome is a 42 bp perfect palindrome. Downstream from the palindrome, an open reading frame (ORF) was identified that revealed strong DNA homology to the telN gene of Escherichia coli phage N15 encoding a protelomerase. Similar to PY54, the N15 prophage is a linear plasmid with telomeres. The N15 protelomerase has cleaving/joining activity generating the telomeres by processing a 56 bp palindrome (telomere resolution site tel RL). To study the activity of the PY54 protein, the telN-like gene was cloned and expressed in E. coli. A 77 kDa protein was obtained and partially purified. The protein was found to process recombinant plasmids containing the 42 bp palindrome. Telomere resolution of plasmids under in vivo conditions was also investigated in Yersinia infected with PY54. Processing required a plasmid containing the palindrome as well as adjacent DNA sequences from the phage including an additional inverted repeat. Regions on the phage genome important for plasmid maintenance were defined by the construction of linear and circular miniplasmid derivatives of pY54, of which the smallest miniplasmid comprises a 4.5 kb DNA fragment of the plasmid prophage.