Transposable element Tcl of Caenorhabditis elegans recognizes specific target sequences for integration ( site specificity / transposon / unc - 22 / nematode )

The frequency of movement of Tcl, a 1.6kilobase transposable element in the nematode Caenorhabditis elegans, is under genetic control, and Tcl insertion sites are widely but nonrandomly distributed. The unusually high frequency of insertions at multiple sites in the gene unc-22 suggested that this gene might be particularly rich in preferred target sites. To discover the features of Tcl target sites, we have sequenced the sites of seven independent Tcl transpositions into unc-22 and three other sites. Our comparison of these and two other sites from the literature indicates that in all cases Tcl integrates at the dinucleotide T-A when it is flanked both 5' and 3' by particular preferred nucleotides. Our analysis revealed the following consensus target for Tcl integration: G-A-K-A-T-A-T-G-T, in which K = G or T. This target site sequence specificity has implications both for the mechanism of Tcl transposition and the use of Tcl in cloning genes by transposon-tagging. A significant fraction of eukaryotic genomes consists of repetitive sequence families. Very often these repetitive sequences have structural features common to transposable elements, which have the ability under appropriate conditions to move in the genome and thereby to cause mutations. The insertion-site specificity is known for several prokaryotic elements (1-4) and has been inferred for some eukaryotic elements (5-8). The features of the insertion site have implications for the mechanism of transposition and possibly for genome evolution of the organisms that harbor the elements. More practically, insertion-site specificity may affect the usefulness of the transposon for cloning genes by insertional mutagenesis (9, 10). In the nematode species Caenorhabditis elegans, the Tcl transposable element was initially identified as a repetitive sequence (11-14). Sequencing of one member showed that the element is 1.6 kilobases (kb) in length, has 54-base-pair (bp) perfect inverted repeats at its ends, and has a large open reading frame (11). This element is present from 30 copies to >300 copies in various strains (13, 14). More recently, a strain of C. elegans called Bergerac or BO was found to exhibit a high frequency of spontaneous unstable mutations (15) caused by the insertion of Tcl (9, 10, 16). The frequency of insertional mutagenesis varies among genes. One of the most frequently targeted genes is unc-22, a muscle-affecting gene (9, 15), where the frequency of insertion is about 10-4. The Tcl insertion sites of 15 unc-22 mutations are scattered into at least 12 different sites, which span a 30-kb stretch of the unc-22 region (17). On the other hand, the mutation frequency of unc-54, a gene encoding the major myosin heavy chain, is <10-' in the same or related backgrounds (16). The difference in gene size alone (about 3-fold) does not explain the difference in mutation frequency. One plausible explanation for the variation of Tcl insertional mutagenesis among genes is that Tcl has a preferred target sequence for insertion and the number of target sequences varies among genes. The wide distribution of Tcl insertion sites in unc-22 suggests that the site specificity for insertion may not be extremely rigorous. The unc-22 gene may also contain more such target sequences than other genes. To test these hypotheses, we have sequenced Tcl insertion sites of 7 unc-22 mutations and 3 other sites unrelated to unc-22. The comparison between these 10 sites and 2 other sites clearly demonstrates that Tcl has a relaxed target site specificity, which probably controls in part the transpositional behavior of Tcl in the Bergerac (BO) background. MATERIALS AND METHODS Cloning and Sequencing of the Tcl Insertion Sites. Nematode strains containing unc-22:: Tcl mutations (st139:: TcJ, stI40::Tcl, stJ41::Tcl, and stJ85::Tcl) were grown for DNA extraction. Total genomic DNAs were purified by the standard methods (18, 19). Genomic DNAs were digested with Sst I and fractionated on agarose gels. The fragments including Tcl-inserted unc-22 fragments were recovered from the gel and ligated into phage A2001 (20). These partial genomic libraries were screened by plaque hybridization (21) with a nick-translated (22) Tcl probe as well as appropriate anc-22 probes, and the desired Sst I fragments containing unc-22:: Tcl insertions were recloned into pUC19 (23). A phage A clone containing the unc-22(st192:: Tc) insertion was provided by J. Kiff in our laboratory. The Sst I fragments containing unc-22(st136:: Tcl) and unc-22(st137:: Tcl) (9), and the HindIll fragment containing Tcl insertion for stPJ, an unc-22-linked dimorphism due to a recent transposition (D.G.M., unpublished data), were previously cloned into pBR322 (24). A clone containing the lin-12 (n137e1979:: Tcl) insertion site was provided by I. Greenwald (10). A clone containing the Tcl insertion in unc-15 was provided by S. Rioux of this laboratory, and the sequence was determined by him as part of these studies. To sequence from the Tcl into the flanking unique sequence, a fragment containing only one Tcl inverted repeat was subcloned into phage vectors M13mpl8 or M13mpl9 (23) in an orientation such that a synthetic oligonucleotide complementary to the Tcl inverted repeat could be used as a primer in dideoxy chain-termination DNA sequencing (25). The M13 universal primer (Amersham) was also used in some cases. Sequences were analyzed by using the Beckman Microgenie program (Beckman). Computer Analysis of the Insertion Sequences. We derived a simple computer program that evaluated all possible strand *Present address: Department of Pathology, Emory University School of Medicine, Atlanta, GA 30322. tPresent address: Department of Zoology, University of British Columbia, BC, Canada V6T 2A9. 861 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact. Proc. Natl. Acad. Sci. USA 85 (1988) comparisons (there are 211 or 2048 for the 12 s find the configuration that gave the best cc quence. For each configuration, the program de overall base preference at each position, fr position to the + 5 position, and then assigned en site a score based on the frequency at which occurred at each position. For example whe preferred base table in Fig. 2 Right derived fo uration shown in Fig. 2 Left, unc-22(st136:: Tc score of 6 for the first position, 6 for the secol third, 4 for the fourth, etc., for a score of 68. 1 each site was summed to give the total s5 configuration of 824. This score was the bes among 2048 strand comparisons. If all 12 sites cal, the score would be (12 points x 10 bases 1440. RESULTS AND DISCUSSION To begin our studies of Tcl insertion, we anal pendent unc-22 mutations (Fig. 1). The sti sti92:: Tcl alleles, though isolated independent inserted at precisely the same site and, as de restriction mapping, in the same orientation. Be unc-22 alleles, we also sequenced an insert in Ii unc-22-linked dimorphism, stPJ (D.G.M., data), and an insertion into unc-15, the structi paramyosin (ref. 26; H. Kagawa, personal comn The sequences surrounding these 10 Tcl inst with two sites sequenced by Rosenzweig et al shown in Fig. 2 Left. We have searched these sequences for comn At each of the 12 sites,Tcl is inserted at a T-A As indicated previously (11, 12), this T-A may 1 upon insertion and is presumably an obligate insertion site. When the flanking sequences we certain preferred bases at the same positions re Tcl insert were obvious. However, the palind of the T-A dinucleotide, and in some cases t flanking sequences [e.g., A-T-G-T-A-C-A-T and A-T-A-T (sti 92:: Tcl)] introduced an ambij analysis-namely, that either the top or botto