T I G - - A u g u s t 1987, Vol. 3, no. 8
The transposable genetic element Tcl in the nematode
At the most recent Caenorhabditis elegans meeting (held at Cold Spring Harbor Laboratory, May 6-10, 1987), nearly 30% of the 188 meeting abstracts referred to a C. elegans transposable genetic element called Tcl. One reason for the Robert K. Herman and Jocelyn E. Shaw great interest in Tcl has to do with the important questions Tcl is a 1.6 kbp DATA sequence present in about 30 copies in some strains of C. raised by transposable eleelegans and 300 or more copies in other strains. Tcl elements excise much more ments in general, such as how frequently tn somaffc cells than in the germ line. Germ.line transposition of Tcl has they transpose, how they afbeen detected and is under genetic control. Tcl has become very useful as a tool for fect gene expression, and what cloning C. elegans genes zdentified solely by mutation. kinds of mutations they promote. Most C. elegans workers, however, are interested in T c l for a second element (Tcl) present in about 30 copies in N2 and reason: to use it as a tool for cloning genes that have 300 or more copies in BO (Refs 13 and 14). There was early evidence for frequent Tcl excision been identified by mutation and for which no gene in somatic cells, but definitive evidence for germ-line products are known. Study of the developmental genetics of C. elegans, transposition came somewhat later. Moerman and which was pioneered by Sydney Brenner ~, is Waterston discovered that the frequency of spontanbenefiting from the exceptionally detailed description eous mutation of the gene uric-22 was much higher in now available of wild-type C. elegans development. BO than in N2 and suggested that the higher mutability The lineages of all cells generated during development of BO might be due to Tcl transposition 15. Eide and are knownz-4, and the structure and connectivity of Anderson made use of a special screen developed by the complete net ,,)us system have been determineds-s. Park and Horvitz16 to identify a large number of The developmental abnormalities of many mutants spontaneous mutations in the myosin heavy chain gene affected in cell lineage, nervous system development, uric-54, which had already been cloned17; they found muscle development or sex determination have been that a large fraction of spontaneous mutations in BO precisely described with reference to wild-type (but virtually none in N2) were insertions of Tcl (Refs development. Hundreds of genes affecting these 18 and 19). The genes unc-22 (Ref. 20) and lin-12 aspects of development have been identified by (Ref. 21) were the ~ s t of the genes previously analysis of mutants, but for most of them, gene described only genetically to be tagged and cloned by products are unknown. There is thus great interest in Tcl transposition. cloning these genes to examine their gene products; in addition, transformation can now be used to see how Structures of Tc 1 e l e m e n t s and target sites for cloned genes and variants of them function in living Tcl insertion animals~. Rosenzweig, Liao and Hirsh22 compared the nncTcl is contributing to the cloning effort in two ways. leotide sequence of a Tcl-filled site in BO with the First, there are C. elegans strains with hundreds more sequence of the empty (and presumably never occucopies of Tcl distributed throughout their genomes pied) site in N2. In place of the 2 bp sequence TA in than are found in the wild-type strain N2, on which N2, BO had 1614 bp: TA followed by 1610 bp followed most genetic and developmental work has been done. by TA (Fig. 1). One interpretation of this finding is that Each Tcl dimorphism (nresence or absence of Tcl at there was a target duplication of TA upon insertion of a specific site) can be mapped genetically, and the Tcl, in which case Tcl is 1610 bp long. Target site genomic sequence flanking the Tcl can be isolated. duplication ha~ been found for other transposable Each mapped Tcl site is thus a signpost from which elements and is thought to arise as a consequence of a chromosomal walks can be initiated. Such walks are staggered cut at the target site during transposition2a. greatly aided by the availability of many groups of An alternative possibility in the case of Tcl is that the clones known to overlap in nucleotide sequence element is 1612 bp long and is inserted between the T (contigs), which have been generated in a project and A of the target site without target site designed to construct a complete physical map of the duplication 24 (Fig. 1). The sequencing of an additional C. elegans genome l°. The second use for T c l is in the 16 target sites has not resolved this ambiguity because direct tagging by insertional mutagenesis of genes to they all have TA at the insertion site. be cloned, as has been done with other transposable The 1610 bp sequence has perfect 54 bp inverted elements in other organisms (e.g. Refs 11 and 12). repeats at its ends. It also contains two long open reading frames on the same DNA strand. The larger Identification of Tc 1 open reading frame corresponds to a basic polypeptide Tcl was first identified as a consequence of cf 273 amino acids. TATA and CAAT box sequences Southern blot comparisons made between two C. for transcriptional initiation are present at appropriate elegans strains, Bristol N2 (English strain) and spacings 5' to the putative coding region, and a Bergerac BO (French strain). Unique sequence possible polyadenylation signal was identified. The probes identified restriction fragments that were 1.6 second open reading frame, which potentially encodes kbp larger in BO than in N2, and it was shown that the a ll2-amino acid polypeptide, is entirely within the extra 1.6 kbp was due to the presence of a repetitive first but in a different translational reading frame. No
© 1987, Else'aer Pubhmlaons. Cambridge
0168 - 9525/87/$02.00
August 1987, Vol, 3, n,~. 8
FiE. 1. Mejor features of ~'cl and its i~etti~n target site, basedon n=cl~eti~ seq~ndr_~"z'z4and showing the terminal invert~ repeats (IE) and the largest open r e ~ i ~ ~a~r,~ (ORF).
TATA or CAAT box sequences with appropriate spacings were found for the latter sequence. Possibly the same message is translated in two ways. Structurally, Tcl resembles the P elements of Drosophila, the Ae-Ds elements of maize, and the bacterial IS elements; and as is the case for these elements, Tcl transcripts are not abtmdant. Unlike P elements and Ac-Ds elements, for which abbreviated variants are common, nearly all the T c l copies seem to be the same size13"t4. Slight heterogeneities in restriction enzyme sites have been noted within the Tcl family, howevert°'~s. The insertion of Tcl shows a bias for specific target sequences. Eleven independent Tcl insertions in uric54 were found to have occurred at just four sites (at the nucleotide level), all in exons (D. Eide and P. Anderson*). [By contrast, in uric-22, which is clearly a favored gene for Tcl insertion, 12 Tcl insertioas occurred in at least ten separate sites, spanning about 30 kbp (Ref. 20)]. Inspection of 17 target sequences that have been analysed shows that the only absolutely conserved sequence is the TA insertion site. The sequences flanking the TA targets are not random, however, and a 9 bp consensus target sequence has been proposed (I. Mori, G. Benian, D. Moerman, and R. Waterston, pets. commun.).
Tcl excision: somatic versus germ-line High frequency excision of the Tcl element from a particular genomic site can be recognized on a Southern blot. Unique sequence flanking several Tcl elements has been cloned and used to probe Southern blots of genomic DNA. In each case the probe hybridizes to two bands: a heavy band corresponding to the Tcl-bearing fragment and a weak band, 1.6 kbp smaller, corresponding to the Tcl empty site. Such empty sites seem to be generated by most, if not all, Tcl elements in BO (Refs 26, 19, 20). The fact that the empty site is 1.6 kbp smaller suggests that excision is often precise or nearly precise. The empty sites are produced almost exclusively by somatic cells, since they are passed on to progeny at relatively low frequency2~. As a consequence of the low frequency germ-line excision, the fraction of sites that axe empty * Names cited without a reference refer to presentations made at the Cold Spring Harbor C. elegans meeting.
is low iv_ embryos, increases during larval development, and then drops as the germ line becomes more prominent in the adultzT. The much higher excision of T c l in somatic cells than in the germ line is just the opposite of the behavior of P elements in Drosophila during hybrid dysgenesis: the excision and transposition of P elements ;~ germ-line specific (owing to a germ-line specificity in the splicing of the P element transcript) ~. Emmons and colleagues have measured the rate of formation of empty sites at five different genomic sites in BO worms arrested as first stage larvae by starvation in buffer26. At all five sites excision continued for the two weeks that the worms remained alive, by which time more than 10% of sites had been vacated. Thus excision does not seem to require DNA replication. At two of the five sites, N2 also has T c l elements; both showed somatic excision but at a lower rate than in BO. Somatic excision of T c l from within uric-,54 has been followed by the method just described and also by the identification of somatic mosaics, uric-54 mutants are both paralysed and unable to lay eggs because their muscles are defective. Eide and Anderson found that a BO strain carrying a T c l element in uric-54 gave about 1% egg laying-proficient animals TM. These animals usually also showed improved ability to move, and when their longitudinal strips of body muscle cells were inspected by polarized light microscopy, they were often found to have patches of birefringence characteristic of wild-type body muscle (although not fully wild-type). The frequency of germ-line reversion of the same Tcl element was about 4 x 10-6; by contrast, empty sites reached a level of 1-5%. K-S. Ruan and S. Emmons (Ref. 29 and pers. commun.) and D. Eide and P. Anderson have cloned and sequenced several vacated sites generated by somatic excision of Tcl elements at four different locations; most of the excisions were imprecise, usually leaving behind 4 bp insertions, but all of the six excisions examined at one particular site were precise, suggesting that flanking sequence may influence the accuracy of excision. Germ-line T c l excisions from two sites in unc-54 (D. Eide and P. Anderson) and one site in uric-22 (.1.Kiff, D. Moerman and R. Waterson) have been studied by sequencing revertants. Despite the selection imposed for
TIG m August 1987, VoL 3, no. 8
x N2 .~
Self and r ::> select m/m
0 Repeal prev=ous two steps many tumes N2
SOUTHERN BLOT PROBED WITH Tc 1
which the fraction of empty Tcl sites did increase; thus, if the extrachromosomal copies are the product of somatic excision then they must he unstable.
Y m m
Fig. 2. Scheme for Tel-tagging a C. elegans gene to be clonedz°'zl. The BO strain contains about300 or more copies of Tcl, and N2 contains about 30 stable copies (the illustrated Southern blot showing five bands for N2 is ~,aplified for diagrammatic purposes). A modification of this scheme is to substitute a low-copy number Tcl-transposing strain for BO (see text). Caenorhabditis elegans hermaphrodites have five pairs of autosomes (l-V) and a pair of X chromosomes.
restored gene function, the great majority of sequenced germ-line excisions were imprecise. Most revertants at one unc-54 site left behind the same 4 bp insertion often seen after somatic excision; the Tcl insertion site was in an exon but very near an exonintron border, and the extra 4 bp left by the excision appears to have shifted the splice site by 4 bp, with the consequence that a wild-type polypeptide was generated. Reversions at the other sites included short in-frame insertions, duplications of the sequences flanking the insertion site, and 1-2 kbp deletions. Thus, although Tcl excision in BO is much less frequent in the germ line than in the somatic cells, the spectra of vacated sites in the two cell classes appear to be similar if not identical. In conjunction with the investigations into Tcl excision in vivo, an in vitro system desig~led to monitor Tcl excision (R. Plasterk) should help to elucidate the mechanisms of Tcl movement. Extraehromosomal copies of Tc 1 Extrachromosomal copies of Tcl have been identified in BO (Refs 30 and 31), at a level between 0.1 and 1.0 copy per cell. The predominant form was a 1.6 kbp linear molecule with ends corresponding to the ends of the integrated Tcl. Also identified were relaxed and supercniled circular copies of the element. It has been suggested that the extrachromosomal copies of Tcl are the product of somatic excision and may be intermediates in the transposition process~ However, the concentration of extrachromosomal copies did not increase in starved BO larvae2s, in
Although no C. elegans strain devoid of T c l elements has been found, different strains exhibit very different frequencies of germ-line Tcl transposition, and it is important to understand what is responsible for the variation. Tcl copy number is dearly not solely responsible. Transposition, whether measured by insertion in uric-54 or uric-22, is much greater for BO and TR679 (see below), high-copy number strains, than for N2, a low-copy number strain; but DH424 (isolated in California) is a high-copy number strain and shows little if any Tcl transposition]s'xs. Moennan and Waterston also found that two other Bergerac lines (FR and BL1), both high-copy number strains, failed to show the high frequency unc-22 mutation (later shown to be largely due to Tcl insertion) characteristic of the Bergerac BO strain is. Conversely, Tcl-transposing strains having only about 45-60 copies of Tcl have been generated from mutator/N2 hybrids (M. Finney and R. Horvitz; I. Mori, D. Moerman and R. Waterston), and an N2 strain with mutator activity for uric-22 and increased T c l copy number has been discovered (C. Trent andJ. Hodgldn, pers. commun.). The uric-22 mutator activity in BO was shown to depend on genetic background, not on the uric-22 region itself: an uric-22 region from N2 became mutable when put in a BO background, and a BO unc22 region was stabilized in an N2 background~~. Early attempts to localize genetically mutators responsible for Tcl transposition in BO indicated .that they were polygenlc. More recently, mutators in different derivatives of BO/N2 hybrids have been mapped to three linkage groups, and it has been suggested that the mutators may themselves transpose (I. Mori, D. Moerman and R. Waterston). P. Anderson and colleagues have identified strains following EMS mutagenesis that show enhanced frequencies of spontaneous Tcl excision and transposition in the germ line. One of these strains, called TR679, and some of its derivatives have been used for transposon tagging (see below). Portions of the mutator activity of TR679 have been mapped to different linkage groups; the enhanced mutator activity of TR679 may be polygenic or transposable (or both). The mutator activities in both BO and TR679 appear to be dominant to inactivity. Mutators could be providing trans-acting factors that promote transposition of Tcl elements in the genome; if so, they might be analogous to a maize Ac element or a Drosophila P element, each of which is capable of mobilizing non-autonomous elements, or they might be acting in some other way. An apparent distinction between BO and TR679 has been revealed recently by the finding that transposable elements distinct from Tcl are activated in TR679. One such element is called Tc3; it is not homologous to Tcl and has been found in three distinct sites in spontaneous unc-22 mutants of TR679 0. Collins, B. Forbes and P. Anderson) and in two distinct sites in uric-86 mutants recovered in a strain descended from a TR679/N2 hybrid (M. Finney
Auglcst 1987, VoL 3, no. 8
and R. Horvitz). Work on transposable elements in C. elegans other than T c l is in its early stages.
4 Sulston, J. E., Schierenberg, E., White, J. G. and Thomson, J. N. (1983)D~. Biol. 100, 64-119 5 Ward, S., Thomson, N., White, J. G. and Brenner, S. (1975) J. Comp. Neurol. 160, 313-,338 6 Ware, R.W., Clark, D., Crossland, K. and Russell, R. L. (1975) ]. Comp. NeuroL 162, 71-110 7 Albertson, D. G. and Thomson, J. N. (1976)Philos. Trans. R. Soc. London. Ser. B 275, 299--325 8 White, J. G., Southgate, E., Thomson, J. N. and Brenner, S. (1986) Philos. Trans. R. S~. Lr~don. Set. B 314, 1-340 9 Fh-e, A. (1986)EMBO ]. 5, 2673--2680 10 Coulson,A., Sulstan, ]., Brannet, S. and Karn,J. (1986)Prec. Nail Acad. Sci. USA 83, 7821-7825 11 Bingham, P. M., Levis, R. and Rubin, G. IVL(1981) Cell 25, 693-704 12 Federoff, N. Y., Furtek, D. B. and Nelson, O. E. (1984)Prec. Nail Acad. Sol. USA 81, 3825-3829 13 Emmnns, S. W., Yesner, L., Ruan, K. and Katzenber8, D. (1983) Cell 32, 55-65 14 Liso, L. W., Rosenzweig, B. and Hirsh, D. (1983) Proc. Nail Acad. Sci. USA 80, 3585-3589 15 Moerman, D. G. and Waterstnn, P,- H. (1984) Genetics 108, 859-877 16 Park, E-C. and Horvitz, H. R. (1986) Ge~e6cs 113, 853-867 17 MacLeod, A. R., Karn, J. and Brenner, S. (1981) Nature 291, 386-390 18 Eide, D. and Anderson, P. (1985) Genetics 109, 67-79 19 E~de,D. and Anderson, P. (1985) Proc. Nail Aead. Sci. USA 82, 1756--1760 20 Moerman, D. G., Benian, G. M. and Waterston, R. H. (1986) Proc. Nail Aead. Sc~. lISA 83, 2579-2583 21 Greenwa]d, I. (1985) Cell 43, 583-590 22 Rosenzweig, B., Liao, L. W. and Hirsh, D. (1983) Nucleic Acids Res. 11, 4201-4209 23 Grindley, N. D. F. and Sherratt, D.J. (1983) Cc/d Spying Harbor Symp. QuanI. Biol. 43, I257-1261 24 Rosergweig, B., Liao, L. W. and Hirsh, D. (1983) Nucleic Ac/ds Res. 11, 7137-7140 25 Rose, A. M., H~x'is, L.J., Mawji, N. R. and Morris, W. J. (1985) Can. ]. Bioci~,m. Cell Biol. 63, 752--756 26 Emmons, S. W., Roberts, S. and Ruan, K. (1986) Mol. Gen. Genet. 202, 410-415 27 Emmons, S. and Yesner, L. (1984) Cell 36, 599-605 28 Lanlu, F. A., Rio, D. C. and Ruhin, G. M. (1986)Ce1144, 7-19 29 Emmorm,S. W., Ruan, K-S., Levitt, A. and Yesner, L. (1985) Cold Spring Harbor Syr~. Quant. BtoL 50, 313-320 30 Ruan, K-S. and Emmons, S. W. (1984) Proc. Nail Acad. Sci. USA 81, 4018--4022 31 Rose, A. M. and Snntch, T. P. (1984) Nature 311, 485-486
Transposon tagging and gone cloning In BO, the frequencies of spontaneous Tcl-induced mutants for uric-54, uric-22 and lin-12 are about 5 x 10 -7, 10 -4 and 5 × 10 - s respectively 19-21. F o r each of these loci, selective methods for identifying r a r e mutant individuals were used. For loci for which mutant selection is not possible, the frequency of T c l insertion may be too low to make BO useful for transposon tagging. Enhanced mutator strains such as TR679 have been used successfully by a number of workers to tag genes of interest. In TR679, the frequency of T e l mutation in uric-22 is about 10 -3 (P. Anderson). Other genes have shown lower but still useful frequencies of T e l insertion. Still others have been refractory to T c l insertion, but at least one of these, uric-86, has been a target for other transposons (see above). Of course, finding a spontaneous mutant for a g e n e of interest in a Tel-transposing strain is only the first s t e p in cloning the gene. One must pick out the T c l tagged sequence from among all the T c l - b e a r i n g sequences in the mutant. T h e now-standard method for identifying the relevant T c l is to outcross the mutant to the wild-type strain N2, to pick homozygous mutant self progeny, and then to repeat these two s t e p s many times 2°'2x (Fig. 2). The chromosomes of the resulting strain should all be N2 except in the vicinity of the spontaneous mutation. Southern blots of DNAs from the mutant strain and from N2 are compared to s e e if an extra Tcl-containing band can be found. Closely-linked T c l elements may lead to extra bands that are not actually within the gene to be cloned; t h e s e may be eliminated by selection for recombination in the intervals on each side of the gene of interest, using genetically marked N2 strains. Once one identifies a Tel-containing fragment that invariably cosegregates with the mutation of interest, fragments of appropriate size can be cloned in E . coli, and a clone containing T c t can be R. K. Herman and J. E. Shaw are at the DepaTlmeut of Genetics and Cell Biology, Dniversity of Minnesota, S t Paul identified. Sequence flanking the T e l can be used for MN 55108, USA. subsequent probing and r e c o v e r y of more flanking DNA. An alternative to Tel-tagging, as already noted, is Reviews scheduled for forthcoming to make use of T e l dimorphisms known to be closely linke;! to the gene of interest to initiate a chromosomal issues of Trends in Genetics walk through the gene (G. Ruvkun, V. Amhros and R. The evolution of mammalian sex chromosomes Horvitz). When it is thought that the desired DNA has and dosage compensation- clues from marsupials been isolated, by either method, additional mutant and monotrsmes, alleles (including revertants of Tcl-induced mutaby Jennifer A. Marshall Graves tions) that affect patterns of Sou t h e m blots can provide Molecular time scale for evolution, confirming evidence that the region of interest has in by Allan C. Wilson, Howard Ochman and Ellen fact been cloned. In addition, transformation with the M. Prater isolated wild-type genes has been reported to rescue Enhancers of site-specific recombination, mutants, providing good evidence that the correct by Reid C. Johnson and Melvin I. Simon DNA sequences were cloned (A. Fire and R. Cell lineages in leech embryogenesis, Waterson; J. Way and M. Chaifie). These strategies by Marly Shankland have already been successful for several genes. A gone conversion program during the Improvements seem likely as more is learned about ontogenesis of chicken B cells, T e l and other transposable elements in C. e l e g a ~ . by Claude-Agn~s Reynaud, Auriel Dahan and Jean.Claude We#tReferences Molecular approaches to the study of gone l Brenner, S. (1974) Eenet/cs 77, 71-94 mutation in human ceils, 2 Sulstan, J. E. and Horvitz, ]I. R. (1977) Dee. B=oi. 56, 110-156 3 Kimble,J. and Hirsh, D. (1979) Dev. Bml. 70, 396-417
by Robert B. DuBfidge and Michele P. C.alos