key: cord-0004716-cer5ifac
authors: Astua-Monge, G.; Lyznik, A.; Jones, V.; Mackenzie, S. A.; Vallejos, C. E.
title: Evidence for a prokaryotic insertion-sequence contamination in eukaryotic sequences registered in different databases
date: 2002
journal: Theor Appl Genet
DOI: 10.1007/s001220200005
sha: 4f49a87e4eecdcaaeb21a47eb3f351cc71059eb5
doc_id: 4716
cord_uid: cer5ifac

An insertion-sequence of prokaryotic origin was detected in a genomic clone obtained from a Phaseolus vulgaris bacterial artificial chromosome (BAC) library. This BAC clone, characterized as part of a contig constructed near a virus resistance gene, exhibited restriction fragment length polymorphism with an overlapping clone of the contig. Restriction analysis of DNA obtained from individual colonies of the stock culture indicated the presence of a mixed population of wild-type and insertional mutants. Sequence analysis of both members of the population revealed the presence of IS10R, an insertion-sequence from Escherichia coli. A BLAST search for IS10-like sequences detected unexpected homologies with a large number of eukaryotic sequences from Homo sapiens, Arabidopsis thaliana, Drosophila melanogaster and Caenorhabditis elegans. Southern analysis of a random sample of BAC clones failed to detect IS10 in the BAC DNA. However, prolonged sub-culturing of a set of 15 clones resulted in transposition into the BAC DNA. Eventually, all cultures acquired a 2.3-kb fragment that hybridized strongly with IS10. Sequence analysis revealed the presence of a preferred site for transposition in the BAC vector. These results indicate that a large number, if not all, of the BAC libraries from different organisms are contaminated with IS10R. The source of this element has been identified as the DH10B strain of E. coli used as the host for BAC libraries.

Nucleic acid databases have become a powerful tool for gene discovery and comparative genomics. Unfortunately, a relatively high proportion of entries have been found to be contaminated with extraneous sequences. While some sources of contamination are endogenous, e.g., ribosomal RNA sequences (Gonzalez and Sylvester 1997) , the most common sources are plasmid vector sequences, bacteriophage M13 (Miller et al. 1999; Seluja et al. 1999) , and sequences from the cloning host (Binns 1993; Jha and Sarkar 1997) . Escherichia coli insertionsequences (IS) have also been found as contaminants of cloned eukaryotic sequences (Binns et al. 1986; Binns 1993) .

IS1 and IS421 were reported to be the most common IS elements contaminating database entries (Binns 1993) . In fact, a recent homology search detected eight eukaryotic sequences contaminated with IS1 and 13 with IS421. However, in recent years there has not been any new reports of additional IS sequences as contaminants of eukaryotic sequences. We report here the detection of a prokaryotic transposable element in the insert of a BAC clone of Phaseolus vulgaris (the common bean). This element is also present in several entries of common databases, including several from the human genome.

Extraction of BAC DNA BAC clones were obtained from a P. vulgaris BAC library (Vanhouten and Mackenzie 1999) . E. coli carrying the BAC clones was grown in Luria-Bertani (LB) medium (Sambrook et al. 1989) containing 34 mg/ml of chloramphenicol. Cultures were grown for 16 h at 37°C with continuous agitation at 250 rpm. BAC DNA was extracted using a BAC DNA purification kit from Pinceton Separations, Incorporated (Adelphia, N.J.), or a modified alkaline-lysis method (Sambrook et al. 1989 ). In the latter method, bacterial cells (2 ml) were pelleted by centrifugation at 1500×g max for 15 min, resuspended in 200 µl of GTE buffer (50 mM Glucose, 25 mM Tris.HCl pH 7.8, 10 mM Na EDTA) and lysed by the addi-tion of 300 µl of a solution containing 0.2 M NaOH/1% SDS at room temperature for no more than 5 min. The mixture was neutralized for 5 min on ice with 300 µl of 5 M potassium acetate (pH 4.5). The lysate was centrifuged for 10 min at 21000 g max and the resulting supernatant was transferred to a new tube for extraction with 1 vol of chloroform. After inverting the tube by hand 4 to 5 times, the phases were separated at 21000 g max for 5 min. The upper phase was mixed with 1 vol of isopropanol and the DNA was pelleted at 21000 g max for 30 min. The pellet was washed in 500 µl of 100% ethanol, then dried and resuspended in 30 µl of 0.1× TE buffer. In contrast to results obtained with the purification kit, alkaline lysis preparations tended to have some chromosomal DNA contamination.

BAC DNA was digested with EcoRI or EcoRV according to the manufacturer's instructions (Life Technologies, Grand Island, N.Y.) . DNA restriction fragments were resolved by agarose-gel electrophoresis (Sambrook et al. 1989 ) and transferred to a Biodyne nylon membrane (Life Technologies, Grand Island, N.Y.) as described by Southern (1975) and modified by Chomczynsky (1992) . The probes were labeled with 32 P using a Random Primed DNA Labeling kit (Boehringer Mannheim, Indianapolis, Ind.), or PCR (Plyler and Vallejos, 2000) . Primers for PCR-amplification of IS10R were designed with Primer3 (http://www_genome. wi.mit.edu/cgi_bin/primer/primer3.cgi). These primers [(5′CCTC-ATAATTTCCCCAAAG3′) and (5′GCGAATGCAGATTGAAGA-AAC3′)] amplified a 351-bp product. Hybridization signals were visualized by autoradiography on X-OMAT film (Kodak).

Sequencing was carried out at the Interdisciplinary Center for Biotechnology Research (ICBR) Sequencing Core Lab (University of Florida, Gainesville, Fla.) and the Iowa State University DNA Sequence Facility (Ames, Iowa) using the Applied Biosystems model 373 or 377 Prism system (Applied Biosystems, Foster City, Calif.). The computer programs SeqAid II version 3.81 (Shareware by D. D. Rhoads and D. J Roufa, Kansas State University) and BLAST 2.0 (Altschul et al. 1997 ) were used to analyze nucleotide sequence data and predicted protein products.

We are using a P. vulgaris (the common bean) BAC library (Vanhouten and Mackenzie 1999) to construct a contig around a virus resistance gene. For the purpose of this report we focus on two overlapping clones of the contig: BnBAC25 and BnBAC55. Restriction fragments containing the termini of BnBAC25 were isolated and cloned previously (Plyler and Vallejos 2000) . BAC25L was identified as containing one of the termini of BnBAC25 and also as one of the leading ends of the contig. Southern hybridizations of this clone to genomic blots of P. vulgaris DNA digested with six restriction enzymes indicated that this clone represents a single-copy genomic sequence. BAC25L was used to screen the P. vulgaris BAC library and isolate BnBAC55. Overlap between these clones was inferred by analysis of restriction fragment profiles and by hybridization of Southern blots with 32 P-labeled BAC25L. Southern hybridization of EcoRI-digested BnBAC25 and BnBAC55 clones vulgaris BAC clones resulting from the insertion of the prokaryotic transposable element IS10R. The numbers on each lane in panels A and C represent the BnBAC clone designation, and those on panel B the colony number of BnBAC55. A A southern-blot probed with BAC25L showing that BnBAC 25 and BnBAC55 share a 2.8-kb EcoRI fragment, but BnBAC55 possesses an additional 4.1-kb cross-hybridizing fragment. B A southern-blot of individual BnBAC55 colonies showing the 2.8-or the 4.1-kb fragment that hybridizes to BAC25L. C A southern-blot of several BAC clones showing that only the mutant BnBAC55::IS10R has a 4.1-kb fragment that hybridizes to a probe derived from the IS10R sequence. D Graphic representation of the results depicted in the other panels probed with BAC25L revealed the presence of a 2.8-kb fragment in both clones. This is exactly what would be expected of overlapping clones because the P. vulgaris BAC library was constructed with partially digested EcoRI fragments.

When the hybridization described above was repeated a few months later with a new Southern blot of the BAC clones, both BnBAC25 and BnBAC55 displayed the same 2.8-kb restriction fragment as before, but BnBAC55 displayed an additional 4.1-kb fragment (Fig. 1A) . Three explanations were considered for the new fragment in BnBAC55: (1) partial digestion, (2) a mutation at one of the EcoRI sites, or (3) an insertion event. The first explanation was discarded because inspection of the EcoRI restriction profile of BnBAC55 showed that the digestion had been carried out to completion. To discriminate between the latter two alternatives, BnBAC55 cells from the original freezer stock were plated on chloramphenicol plates and eight single colonies were selected for BAC DNA extraction and analysis. These DNAs were digested with EcoRI, blotted onto nylon membranes and hybridized with BAC25L. Figure 1B shows that one group of colonies contains the 2.8-kb fragment while the other group contains the 4.1-kb fragment. To identify the nature of the shift in molecular weight, one colony from each group was selected for sequencing in the region encompassing BAC25L. Primers derived from the BAC25L sequence were used for this purpose and for sequencing the region adjacent to BAC25L in BnBAC25 as well.

Sequence comparisons among BnBAC25, the BnBAC55 wild-type and the BnBAC55 mutant revealed the presence of a 1329 bp insertion in the mutant. A sequence-homology search of the insertion-sequence with Gapped BLAST 2.0 (Altschul et al. 1997) indicated 100% identity to the IS10 element found in enterobacteria such as E. coli and Shigella flexneri. To our surprise (and chagrin) many other sequences in the database from diverse sources such as Homo sapiens, Drosophila, Arabidopsis, and Picea abies chloroplast DNA contain IS10 (Table 1) . Alignment of this sequence with the two inverted repeats of the complex Tn10 transposable element indicated that this sequence was identical to the 3′ end repeat of Tn10 and, therefore, was identified as IS10R. The alignment also identified the expected 9-bp direct duplication flanking the insertion. The observed duplication contained a deviation from the reported consensus sequence for insertion: Consensus 5′nGCTnAGCn3′ BnBAC55::IS10 5′aGCTaTGCt3′.

Deviations from the consensus have been observed at each position by others (Kleckner et al. 1996) . Inspection of the sequences listed in Table 1 also revealed the presence of the 9-bp direct repeat with variants at every position. Figure 1C shows that only the mutant BnBAC55:: IS10R has a 4.1-kb fragment that hybridizes with the IS10R probe. A weakly hybridizing fragment can be detected in the BnBAC25 lane in Fig. 1C . This signal could be due to contaminating E. coli chromosomal DNA in the BAC DNA preparation, or to an incipient mutant population carrying a newly inserted copy of IS10R. Figure 1D presents a model to explain the results observed in panels A, B and C. BnBAC25 and BnBAC55 are overlapping clones of a contig. A 2.8-kb EcoRI fragment that contains BAC25L is present in the overlapping clones. The insertion of IS10R, which lacks an EcoRI site, is postulated to have occurred approximately 300 bp from the first EcoRI site downstream from the BAC25L sequence. This insertion causes the formation of a 4.1-kb EcoRI fragment that hybridizes with both BAC25L and the truncated IS10R probe.

To investigate the source of IS10R, we analyzed the E. coli strain DH10B, host for the P. vulgaris BAC library. Cells of this strain were grown from the same freezer stock (electrocompetent cells obtained from the manufacturer) used for the construction of the library, and genomic DNA was isolated for PCR and Southern-hybridization analysis. PCR-amplification was carried out with nested primers derived from the sequence obtained from BnBAC55::IS10R. A clear PCR amplification product of the predicted truncated size was obtained in these assays. Furthermore, Southern-blot hybridization of DH10B genomic DNA with the truncated IS10R amplification product revealed several hybridizing fragments. These results indicated that DH10B was the source of IS10R. This observation prompted us to investigate whether DH10B was harboring the typical Tn10 transposon that carries the tetracycline resistance gene. To test this hypothesis, cells of BnBAC25, BnBAC55 and BnBAC55:: IS10R were grown in LB containing 25 µg/ml of chloramphenicol, 6 or 12.5 µg/ml of tetracycline, and combinations of each tetracycline concentration with chloramphenicol. Growth was observed only when tetracycline was absent, suggesting that IS10R alone, and not the typical Tn10, was present in the cells.

The entire BAC library (33000 clones) was screened with the truncated IS10R amplification product. Two levels of hybridization signal intensity were observed (Fig. 2) . Clones exhibiting the strongest hybridization signals appeared to contain IS10R in the BAC DNA, whereas those exhibiting the weaker hybridization signal contained only the genomic copies. BAC DNA isolated from the intensely hybridizing clones, arising in approximately 1/10000 clones, contained a 2.3-kb EcoRV fragment that strongly hybridized to the IS10R probe (data not shown).

To better-characterize transposition in clones of the BAC library, we selected at random 30 BAC clones for analysis after an overnight culture and a subset of these after 8 days of daily subculturing. EcoRV digestions and Southern-blot hybridization with the truncated IS10R probe clearly showed the presence of five distinct fragments in each one of the 30 samples. These fragments were deemed to contain IS10R copies of genomic origin because they were identical to those obtained with genomic DNA from untransformed DH10B cells. These results are consistent with the observation that BAC DNA isolated with the miniprep procedure contained small amounts of chromosomal DNA contamination. Additional bands were also observed and probably represented recent transposition events. Figure 3 A shows the results for 15 of the 30 BAC DNA preparations. Ten of these clones were immediately subcultured on a daily basis for a period of 8 days. At the end of this period BAC DNAs were analyzed to determine whether IS10R had transposed. Eight of these samples are shown in Fig. 3B . Southernblot hybridizations with the subcultured clones revealed the disappearance of some fragments (compare clones 4 and 7 from Fig. 3A and B) , and the appearance of several new bands of varying intensity. Among these, a 2.3-kb EcoRV hybridizing fragment was noticed in each clone. This new fragment co-migrated with that contained in the BAC DNA isolated from the clone (30G19) exhibiting strong hybridization to IS10R in Fig. 2 .

The presence of the 2.3-kb EcoRV fragment in all clones after prolonged subculturing suggested that IS10R Fig. 2 Sample BAC filter-hybridization results with the IS10R sequence as a probe. Intensely hybridizing clones are observed with a frequency of 10 −4 . BAC DNA prepared from the intensely hybridizing clones possesses a common restriction fragment that contains part of the IS10R sequence database sequence entries from diverse eukaryotic sources that include humans, Arabidopsis, yeast and Drosophila, among others. The natural presence of a highly conserved prokaryotic transposable element in diverse sequences of eukaryotic organisms seems highly unlikely and is believed to be a contaminant originating from the E. coli strain DH10B. The evidence to support this assertion is based on the observation that the majority of the IS10R-contaminated eukaryotic sequences detected in the databases were from BAC clones, and the fact that the E. coli DH10B strain is commonly used as a host for BAC libraries. However, more-definitive evidence was provided by hybridization and PCR-based detection of this element in the DH10B stock (obtained from Life Technologies, Inc) used as a host for the P. vulgaris BAC library. The detection of 9-bp direct repeats flanking IS10R in the database entries removes the possibility that sequence contamination could have occurred in silico. In some instances, the contaminated sequences have been annotated with regards to the presence of IS10, but the contaminating sequence has not been filtered out. This indicates that, although some researchers are aware of the contamination, the extent and origin of IS10R contamination is not widely known in the scientific community. To our knowledge, this contamination has not been reported in the literature.

Tn10 is a complex prokaryotic transposon comprising two inverted repeats of IS10 flanking a tetracycline resistance gene . Sequence analysis indicated that the IS element detected in BnBAC55 corresponds to IS10R, the copy present at the 3′ end of Tn10. This copy is an autonomous mobile element because, in contrast to IS10L, it has a functional transposase . The analysis presented here suggests that IS10R alone is present in the E. coli DH10B strain used in the construction of the P. vulgaris BAC library. This is probably true for the other BAC libraries. The presence of multiple hybridizing fragments on a Southern blot of genomic DNA of untransformed DH10B can be ascribed to the presence of multiple copies of the element in the DH10B genome. Eichenbaum and Livneh (1995) observed that transposition of IS10R was usually accompanied by an increase in the number of fragments that hybridized with the IS10R probe. The proposed mechanism of transposition is based on the fact that IS10R transposition is controlled by dam methylation. It is hypothesized that excision occurs in one of the hemimethylated sister molecules soon after the replication fork passes the IS10 element. The donor molecule is degraded, and an insertion may occur at a point where the replication fork has not yet passed. As a result, the surviving sister molecule will carry two copies of the element .

The experiments presented here showed that IS10R can transpose into BAC DNA (vector or insert) with continuous subculturing. The existence of a preferred site for IS10R-transposition in the BAC vector is strongly suggested by the presence of an IS10R-hybridizing 2.3-kb EcoRV fragment in 30G19, and the appearance of 52 has a preferred site for transposition in the cloning vector. To test this hypothesis, we took two parallel approaches to determine the sequences flanking the transposon from BAC clone 30G19 shown in Fig. 2 . First, the 2.3-kb EcoRV fragment was cloned into the pGEM-5Zf (Promega Corp) plasmid and the ends of the insert were sequenced. Second, IS10R-derived oligonucleotides were used as sequencing primers for the flanking regions. Our analysis revealed that the preferred sequence for transposition corresponds to position 1470 to 1478 of pBeloBAC11 (gi|1817728). This sequence also shows deviations from the consensus sequence:

Consensus 5′nGCTnAGCn3′ BAC 5′cGCCcGGTa3′.

Although three of the six bases deviate from the consensus, the internal four bases are in an inverted-repeat configuration. Transposition at this site does not appear to disrupt any of the vector functions because it is located in an intergenic region upstream from the CM r gene. The 2.3-kb EcoRV fragment includes 265 bp from the 3′ end of IS10R and 2027 bp from the BAC vector.

A prokaryotic element, IS10R, was detected in a P. vulgaris BAC library and, by a homology search, in several a similar fragment in all BAC clones after daily subculture. Transposition of IS10R upstream from CM R in 30G19 appears to have occurred soon after transformation because the population is uniform, whereas the 2.3-kb fragment detected after daily subculturing is not as strong as other hybridizing fragments (Fig. 3B) and is assumed to result from a more-recent transposition event. The strong hybridization signal produced by 30G19 can be explained by a polarity effect (Szeverenyi et al. 1996) . It has been reported that transposition of IS elements upstream from ORFs can bring about increased levels of expression. An increased level of expression of CM R could give a selective advantage and result in increased growth and a higher hybridization signal. At the present time, we do not know whether transpositions into the cloned DNA are intermediary stages before the element ends up at the preferred vector site upstream from the chloramphenicol resistance gene. The presence of IS10R within the genome of the host commonly used for BAC libraries presents serious problems. Perhaps the most obvious of them is the detection of false homologies after transposition into the insert, as indicated earlier. However, this problem can be overcome by appropriately filtering BAC sequences before submission. Another problem can be encountered during fingerprinting of BACs when the mutant population(s) increases to a significant level. More troublesome is the potential problem of alterations in the cloned sequences that can vary from short duplications to deletions and inversions Binns 1993 ) and co-integrate formation in RecA + strains (Einchenbaum and Livneh 1995). We have only monitored the fate of a small sample of clones after continuous daily subculturing for a short period of time. A more detailed analysis of BAC clones is needed before the long-range fate of IS10R-contaminated BAC libraries can be ascertained. BAC clones from libraries known to be contaminated with IS10R can still be utilized as long as some properties of this element are taken into account. The rate of transposition of IS10 has been estimated at 10 −4 per element per generation (Kleckner et al. 1996) , and transposition of this element is negatively regulated by dam methylation (Roberts et al. 1985) . However, continuous incubation at the stationary phase (Skaliter et al. 1992) , as well as exposure to UV radiation (Eichenbaum and Livneh 1998) , can increase the rate of IS10 transposition. Taking into account the properties of the transposable element during the handling of BAC libraries or individual clones may avoid or ameliorate the negative effects of the element. It is obvious that a new host free of transposable elements is needed before the construction of new BAC libraries is considered.

Gapped BLAST and PSI-BLAST: a new generation of protein database search programs

Genetic evidence against intramolecuar rejoining of the donor DNA molecule following IS10 transposition

Contamination of DNA database sequence entries with Escherichia coli insertion sequences

Comparison of the spike precursor sequences of coronavirus IBV strains M41 and 6/82 with that of IBV Beaudette

One-hour downward alkaline capillary transfer for blotting of DNA and RNA

Intermolecular transposition of IS10 causes coupled homologous recombination at the transposition site

UV light induces IS10 transposition in Escherichia coli

Incognito rRNA and rDNA in databases and libraries

A symmetrical six-base pair target site sequence determines Tn10 insertion specificity

DNA sequence organization of IS10-right of Tn10 and comparison with IS10-left

DNA sequencing and comparative sequence analysis reveal that the Escherichia coli genomic DNA may replace the target DNA during molecular cloning: evidence for the erroneous assembly of E. coli DNA into database sequences

Uses of transposons with emphasis on Tn10

Tn10 and IS10 transposition and chromosome rearrangements: mechanism and regulation in vivo and in vitro

A rapid algorithm for sequence database comparisons: application to the identification of vector contamination in the EMBL databases

A method for cloning restriction fragments containing the termini of BAC inserts

IS10 promotes formation of adjacent deletions at low frequency

Establishing a method of vector contamination identification in database sequences

Spontaneous transposition in the bacteriophage λcro gene residing on a plasmid

Detection of specific sequences among DNA fragments separated by gel electrophoresis

Vector for IS element entrapment and functional characterization based on turning on expression of distal promoterless genes

Construction and characterization of a common bean bacterial artificial chromosome library

Acknowledgments We thank Dr. L.C. Hannah and Dr. K. C. Cline for reviewing this manuscript and offering some valuable suggestions. This work was supported in part by USDA grant No. 97121514 to C. E. V. and NIH grant R21 GM54154 and NSF grant IBN 9728329 to S. A. M. Florida Agricultural Experiment Station Publication No. R-07583. All the experiments described here were performed in full compliance with USA laws.