Copyright 0 1991 by the Genetics Society of America
Chromosome Breakage byPairs of Closely Linked Transposable Elements of the Ac-Ds Family in Maize Hugo K. Dooner and Alemu Belachew DNA Plant Technology Corporation, Oakland, Calqornia 94608-1239 Manuscript received November 20, 1990 Accepted for publication August 2, 1991
ABSTRACT Chromosome breaks and hence chromosomal rearrangements often occur in maize stocks harboring transposable elements (TEs), yet it is not clear what types of TE structures promote breakage. We have shown previouslythat chromosomes containing a complex transposon structure consisting of an Ac (Activator) element closely linked in direct orientation to aterminally deleted or fractured Ac CfAc) element have a strong tendency to break during endosperm development. Here we show that pairs of closely linked transposons with intact ends, either two Ac elements-a common product of Ac transposition-or an Ac and a Ds (Dissociation) element, can constitute chromosome-breaking structures, and that the frequency of breakage is inversely related to intertransposon distance. Similar structures may also be implicated in chromosome breaks in other eukaryotic TE systems known to produce chromosomal rearrangements. The present findings are discussed in light of a model of chromosome breakage that is based on the transposition of a partially replicated macrotransposon delimited by the outside ends of the two linked TEs.
T
HOUGH transposable elements (TEs) in maize werefirst detected by MCCLINTOCK (1947, 1948, 1949) because of their ability to break chromosomes, it is only recently that particular TE structures have been implicated in chromosome breakage (RALSTON,ENGLISHand DOONER 1989; DORINGet al. 1989). A common feature of these structures is that they are compound, ie., they contain more than one element and,therefore, multiple T E ends inclose proximity. One structure, containing three TE ends, consists of two closely linked, but recombinationally separable, elements in direct orientation: a 4.6-kilobase (kb) intact Activator (Ac) element and a 2.5-kb fractured Ac (JAc) element that is deleted for the end ENGLISH that lies closest to the intact Ac (RALSTON, and DOONER 1989). Another one,containing four T E ends, consists of two identical copies aof2-kb internal deletion derivative of Ac, arranged so that onecopy is inserted in inverse orientation into the second copy (DORINGet al. 1989). The former structure arose in a bz-m2(Ac) stock. The bz-m2(Ac) allelehas an autonomous (ie., selfmobile) Ac element inserted in the second exon of the bz gene (RALSTON, ENGLISH and DOONER1988) and can mutate t o stable bz alleles as a consequence of Ac transposition (MCCLINTOCK 1956a,b; DOONERand BELACHEW 1989). One such allele, bz-s:2094, carries a 2.5-kb fractured Ac CfAc) element at the previous site of insertion of Ac and, in addition, a transposed Ac(trAc) 0.05 centimorgans (cM) proximal to fAc (DOONER, RALSTONand ENGLISH1988). In lines carrying the bz-s:2094 allele, chromosome Genetics 129: 855-862 (November, 1991)
breaks at or near bz occur with high frequency during the mitotic divisions that give rise to the endosperm. ENGLISHand DOONER We have proposed (RALSTON, 1989) that chromosome breakage is due to the transpositionof a partially replicated macrotransposon (MTn) extending from the distal end of fAc to the proximal end of the trAc (Figure 1) based on the following observations: (1) Breaks occur only if fAc and Ac are closely linked. (2) A deletion carrying an “empty” MTn excision site was recovered and characterized by sequencing the new junction. (3) The fAc and Ac elements are in direct orientation, as wouldbe expected if fAc and Ac provided the left and right termini of the MTn. The key feature of our model is not the singleended fAc element at one end of the MTn, but the presence of opposite left and right transposon ends separated by a stretch of chromosome (the body of the transposon) that is in a different replication state from one or both transposon ends atthe time of transposition. The model predicts that the combination of two homologous transposons that are closely linked and in direct orientation (thus providing oppositeleft and right ends) will causechromosome breaks. Here, we have examined the chromosomebreaking properties of paired transposons, one at a constant location in the bronze ( b z ) locus and the second one at variable genetic distances from bz. MATERIALS AND METHODS Geneticstocksandcrosses: All the stocks used in the present investigation have the genetic background of the
856
fAc
9s
H. K. Dooner and A. Belachew L
e"rn
TABLE 1
R 1
-
I
"
wx
List of derivatives carrying a trAe
(0.05 cM)
FIGURE 1.--Diagram of thecompositetransposonstructure found in the short arm of chronlosome 9 of the bz-s:20Y4 allele. This structure, which can be excised as ;I Ill;1crotrallsposoIl (hlTn) rstentling from the left (L) end of a fAr element to the right ( K ) end of an intact Ar element located 0.05 cM prositnallv, has been post ul;ltect to promote cI1ronwsonIe breaks.
inbred W22 and carry one of the following allelesat the bz locus in the short arm of chromosome 9 (9.9. Bz-McC (purple): the normal progenitor allele of the T E mutations used in this study. bz-mZ(Ac) (bronze-purple variegation): an allele arising from insertion of a 4.6-kb Ac element in the second exon of Bz-McC (MCCLINTOCK1955; RALSTON, ENGLISH and DOONER1988). bz-mZ(DI) (bronze in the absence of Ac; spotted in its presence): the first derivative from bz-mP(Ac); it harbors a 3.3-kb Ds element as a consequence of an internal deletion in Ac (MCCLINTOCK 1962; DOONERet al. 1986). bz-R (bronze): the stable, reference allele at the locus; it is associated with a 340 bp deletion in the transcribed region (RHOADES1952; RALSTON, ENGLISH and DOONER1987). bz-s:2094(fAc) Ac2094 (bronze): a stable derivative of bzm2(Ac) harboring a 2.5-kb terminally deleted or fractured Ac (fAc) element at bz and a trAc at a location 0.05 CM proximal to bz (DOONER,RALSTONand ENGLISH1988). bz-s:2094(fAc) (bronze): a derivative of bz-s:2094(fAc) Ac2094 that has lost the trAc2094 element (RALSTON,ENGLISH and DOONER1989). bz-s:2114(Ac) (bronze): a stable derivative of bz-mZ(Ac) that arose by deletion of 789 bp of bz sequence proximally and RALSTON1988). adjacent to Ac (DOONER, ENGLISH Two classes of chromosomes derived from the above stocks were tested for breakage: (1) bz-mZ(Ac) WAC,chromosomes with two linked Acs, arose directly by transposition of Ac in a bz-mZ(Ac) stock. The increased Ac dosage results in a finely spotted phenotype. (2) bz-mZ(DI) trAc, chromosomes with a Ds element and a linked WAC,arose by recombination between the bz-mZ(DI) reporter allele and a trAc element in experiments designed to map the location of different trAcs relative to bz. To monitor chromosome breaks genetically, the above stocks (all C in constitution) were used to pollinate c Rz ear parents. Kernels where no breaks occur will be purple; those where breaks occur in 9s will display colorless sectors in a purple background (see RESULTS). The kernels from the various testcrosses were examined individually under 7 X magnification with a stereo-zoom microscope, which allowed the resolution of very small colorless sectors (4-8 cells and larger). In general, the size of the colorless sectors seen in kernels carrying two Ac elements was smaller than in kernels carrying one Ac and one Ds element, an observation in agreement with the well documented negative dosage effect of Ac in maize. An increase in Ac dosage results in a delay of the timing and a reduction of the frequency of transposition (MCCLINTOCK 195 1; BRINKand NILAN1952). In each cross, at least 500 kernels were scored for sectors.
RESULTS
Chromosomes having pairs of homologous elements, one at a fixed location in the bz gene and the other one at variable distances away from bz, were
61 I0 7084 6084 2107 6085 7066 61 17 2097 2 106 7080 3137 7068 7067 7072 707 1 7065 7056 21 16 61 14 2094 2094 6087 6058 6067 7082 2103 7074 705 1 6083 708 1 7070 7069 Controls Ac
Ds2 ( D I ) 21 14 fAc
bz-m2 ( A c ) bz-m2 ( A r ) bz-m2 ( D l ) bz-m2 (Dl) bz-m2 ( D I ) bz-m2 ( A r ) bz-m2 ( A r ) bz-m2 (Dl) bz-m2 ( D I ) bz-m2 ( A c ) bz-m2 ( D I ) bz-m2 ( A c ) bz-m2 ( A c ) bz-m2 ( A c ) bz-m2 ( A r ) bz-m2 ( A r ) bz-m2 ( D l ) bz-m2 ( D I ) bz-m2 ( A r ) bz-m2 (Dl) b~-S:2O94( f A c ) bz-m2 ( D I ) bz-m2 ( D I ) bz-m2 ( D I ) bz-m2 ( A r ) bz-m2 ( D I ) bz-m2 ( A c ) bz-m2 ( D I ) bz-m2 ( D I ) bz-m2 ( A c ) bz-m2 ( A c ) bz-m2 ( A c )
+30.8 +27.7
1.4 f 0..5 1.1 f 1 . 1
+23.5 +21.2 +17.5 + I 2.0 +4.7 +3.3 +3. 1 +2.1 + I .4 + I .3 +1.2 + I .2 + I .0
1.5 f 0.6 7.6 f 1.4 1.1 f 1.1 1 .x f 0.7 1.4 f 0.8 24 f 2.0 26 f 2.5 16 f 2.0 17 f 1.3 32 f 2.1 3.3 f 0.8 4.3 f 0.9 38 f 2.7 2.0 f 0.5 2.0 f 0.7 5 1 f 2.3 82 f 1.8 86 f 1 . 1 92 f 1.0 67 f 3.1 83 f 3.2 84 f 3.6 46 f 3.1 87 f 2.3 18 f 1.4 71 f 1.7 56 f 3.2 4 f 2 5.7 f 1.3 1 .0 f 0.7
bz-m2 (Ac) bz-m2 ( D I ) b 1 - ~ : 2 1 1 4( A c ) bz-s:2094 (fAC)
0.0
+0.6 +0.4
+0.3 +0.2 +0.05 +0.05
-0. I -0.2 -0.3 -0.3 -0.5 -0.5 -0.8 -1.1
-2.2 -4.1 -16.9
0.0
5 . 6 f 0.6 0.07 f 0.07 5.3 f 1.2 0.1 a 0.07
a Locations proximalto bz are represented with positive numbers and those distal to bz with negative numbers. * Proportion of kernels with >5 c sectors.
recovered in the course of a study that examined the pattern of transposition of Ac from bz-m2(Ac) to linked sites (DOONERand BELACHEW1989). They fall into two genetic classes: (i) bz-m2(Ac) trAc, chromosomes having two linked Ac elements, and (ii) bz-m2(DI) trAc, chromosomes having a nonautonomous Ds (Dissociation) element at 6 2 and a linked trAc. The 3.3-kb Ds element in bz-m2(DI) is an internal deletion derivative of Ac in the same location and orientation as the Ac element in bz-m2(Ac) (DOONERet al. 1986). The chromosomes with two linked transposons that were tested for breakage are listed in Table 1. T h e y carry a trAc and are identified by the isolation number of the trAc stock. Their bz locus constitution andthe distance between bz and the trAc (DOONERand BELACHEW 1989) are also given in Table 1. We know that in all these chromosomes both ele-
Breakage Chromosome C
C
82
C
h?
Bz
-0
FIGURE2,“Genetic cross used to monitor the loss of the dominant marker C (colored aleurone) located distal to bz in the short arm of chromosome 9 . A c tester is pollinated with C stocks carrying one transposon at bz and anotherone at variable distances from bz. The frequency of colorless sectors in thealeurone provides an estimate of the frequency of breaks that occur during development of the aleurone.
by Transposons
857
linked Acs. All kernels fromcrosses monitoring breakage were scored as having either > o r 5 5 c sectors and the results were expressed in terms of p , the proportion of kernels with >5 c sectors (Table 1).The standard error of p was computed by the formula JP(1 - P)/n. T h e results of the tests for breakage are summarized in Figure 4, where the “efficiency of breakage” of the various chromosomes being tested, expressed as percent of kernels with >5 c sectors, is plotted vs. the genetic distance separating the two transposons, expressed in cM. It is evident that only transposon pairs that are closely linked are capable of producing more chromosome breaks than the single Ac control. Furthermore, among the “breakers” there is an inverse relationship between breakage frequency and inter-transposon distance, so that the higher breakage frequencies (>50%) are given only by combinations of elements separated by 1 cM or less. This can be more clearly seen in Figure 5, which is a close-up of the 5-cM region on either side of bz. Transposed Acs that, in combination with a second element atbz, cause significantly more breaks than the bz-m2(Ac) control are connected with a line in Figure 5 . They map within 3.3 cM from bz. Chromosomes of a bz-m2(DZ) trAc type appeartoproducemore c sectors than chromosomes of a bz-m2(Ac) trAc type. This is to be expected given the negative dosage effect of Ac on transposition frequency observed in maize (McCLINTOCK1951; BRINKand NILAN 1952). Interestingly, not all closely linked transposon pairs behave as chromosome breakers. Of 22 trAcs tested in the interval that lies 3.3 cM on either side of bz, 5 are clearly nonbreakers (Table 1). Assuming that the estimates of genetic distance are reasonably accurate (they are based on populations of >lo00 gametes; DOONER and BELACHEW 1989), these results indicate that a short inter-transposon distance between the transposon at bz and a trAc is nota sufficient conditionfor the transposon pair to cause chromosome breaks.
ments can transpose, which means that the ends of the elements are intact, in contrast to the terminally deleted fAc element of the bz-s:2094 allele. Evidence that the element atbz (either Ac or Ds) can transpose is provided simply by the bronze-purple variegation seen in somatic tissues of plants carrying these chromosomes, as well as by the germinalpurple ( B z ’ ) derivatives that occur regularly among the progenies of such plants. Evidence that the trAc can still transpose, i.e., thatsecondary transpositions of the trAc element occur, was provided by DOONERand BELACHEW (1989). The average frequency of secondary transposition of Ac from its new location in 9s to unlinked sites in the genomewas estimated to be about 5 x 10-3, which is similar to thatof Ac from its original location in the bz-m2(Ac) allele, and low enough not to interfere with the mapping of trAcs. Chromosomebreakswere assayed genetically by monitoring the loss of the dominant marker C (colored kernel, epistatic to b z ) located distal to bz in the same chromosome arm (Figure 2). Normally, kernels from a cross between a c Bz female parent and a C bz male parent will be solid purple. If the C bz chromosome containing the pair of transposons being tested breaks at or around the bz locus during endosperm DISCUSSION development, the distal marker C will be lost resulting in a colorless sector in the kernel. Thus, thefrequency We have tested herethe chromosome-breaking of colorless sectors provides an estimate of the inciproperties of pairs of closely linked transposons bedence of chromosome breaks during the cell divisions longing to the Ac-Ds family of maize TEs. One memthat produce a mature kernel. In crosses with the bzber of the pair, either Ac or Ds, occupied a constant s : 2 0 9 4 allele, which carries fAc and a trAc 0.05 cM location in the bz locus, and the second one, a trAc, apart, almost all the kernels have many (>lo) colorless was inserted in sites located at variable genetic dis(c) sectors of variable sizes, whereas in control crosses tances proximally and distally from bz. We found that with either bz-m2(Ac) or bz-m2(DZ), carrying single in many lines carrying pairsof closely linked transpotransposons, kernels with >5 c sectors are either rare sons, chromosome breaks resulting in the loss of the o r nonexistent (Figure 3). We can explain the presdistal marker C occurred during the development of ence of occasional sectors in bz-m2(Ac), and not in bzthe endosperm and that the frequency of breakage m2(DZ), kernels as being due to the generation by Ac was inversely related to the distance separating the transposition during endosperm development ofchro- t w o transposons. These results support our model of mosome-breaking structures consisting of two closely chromosome breakage (RALSTON,ENGLISH and
858
H. K. Dooner andA. Belachew to/
" "
-t- br-~:2094 ( f k ) Ac2094
bm2(Ac) bmn2(DJJ
0.9
FIGURE3.-Distribution of kernels with colorless sectors incrossesof c testers by C stocks carrying various transposable elements in 9s. bz-n2(Ac) and bzm2(DI) carry single elements, Ac and Ds,respectively, whereas bz-s:2094uAc) Ac2094carry the elementsfAc and Ac, 0.05 cM apart.
No. c sectorwkernel I
I
eo-
60-
FIGURE4,"Relationship between intertransposon distance and chromosome breaks. Chromosome breakage was assayed genetically as shown in Figure 2. The frequency of kernels with >5 c (colorless) sectors provides an estimate of breakage frequency. Locations distal to br are represented with negative numbers and those proximal to br with positive numbers (see Table 1). Stocks carrying a trAc and a Ds at b t are represented with open circles. Stocks carrying atrAc and an Ac at br are represened with solid circles.
40-
wx
cM Distal
cM Proximal
DOONER 1989) that explains breaks as a consequence of the transposition of a partially replicated MTn extending between the outside ends of two closely linked TEs. We previously considered the consequences of transposition of a MTn in which both ends, locatedin different replicons, have begun to replicate, whereas the centerhas not (Figure6 in RALSTON, ENGLISH and DOONER 1989). Both dicentric and acentric chromatids can arise as the consequences of reinsertion of the MTn in inverted orientation relative to the progenitor chromosome. Similar outcomes are also possible if bothends of theMTnare in the same replicon,
separated by a minimum distance, but only a single end has replicated when transposition occurs. We illustrate these outcomes in Figure 6, which is simply an extension of our earlier diagram that depicted a MTn with both ends replicated. Arguments for the requirement of a minimum distance separating the ends of the MTn have already been presented (RALSTON, ENGLISH and DOONER1989). Figure 6a shows a MTn (B-C) in which only one end (B) has replicated at the time of transposition. As in any other transposition, the Ac transposase cuts at three sites: the two MTn ends (AB and CD) and the target o r receptor site (EF). In Figure 6 the receptor
Chromosome Breakageby Transposons
859
FIGURE5.-Close-up of the region immediately adjacent to bz in Figure 4. Distances distal to bz are given negative numbers and those proximal, positive numbers. The lines connect those trAcs that, in combination with an element at br, give significantly more breaks than the single Ac control (bz-m2).
0
0
I
-5
-4
0
”
1
I
I
-3
-2
-1
sh
cM Distal
0
0.
.r
I
I
1
I
P
1
2
3
4
R
bi
cM Proximal
site is shown distal to the MTn, but it can also be located proximal to it. The MTn can reinsert at the EF receptor site in either direct or inverted orientation relative to its original orientation in the chromosome. If the MTn inserts in direct orientation (Figure 6b), the three new junctions are: AD, corresponding to the “empty”site, and EB and CF, corresponding to the new MTn borders. This type of transposition can be most readily visualized as a displacement of the unreplicated D-E fragment, which separates the donor and receptor sites, to the already replicated proximal end of the MTn. Following resolution of the replication forks, the two daughter chromatids carry, respectively, a distally and directly transposed MTn (i) and a deletion of the D-E fragment that was originally adjacent to thedistal end of the MTn (ii). If the MTn inserts in inverted orientation (Figure 6c),thethree new junctions are AD, again corresponding to the “empty” site, and EC and BF, corresponding tothe new MTnborders.This type of transposition sets up chasing replication forks, which have also been postulated in the mechanism of amplification of the 2u plasmid of yeast (MURRAY1987). Resolution of the replication forks would result in a chromatid with a distally transposed MTn (iii) and a dicentric chromatid (iv). However, it is likely that the trailing replication fork, which will not converge with another replication fork, is not resolved during the course of the S phase, leading to the formationof an incompletely duplicated dicentric. If B-C transposed proximally in aninvertedorientation, anacentric chromatid would be formed (not shown). The latter outcome can be visualized by simply shifting the po-
sition of the centromere shown in Figure 6 from the left to the right end of the chromosome, next to the EF receptor site. Chromatid iv in Figure 6c would be an acentricsince it lacks the D-E-F segment thatwould now be adjacent to the centromere.A dicentric chromatid will form an anaphase bridge that will break and initiate a chromatid breakage-fusion-bridge cycle and anacentric will lag at the metaphase plate and be lost from either daughter cell. Though the majority of T E pairs comprising one element at br and a trAc in the vicinity of bz are breakers, some are not. Of 22 trAcs tested in a 6-cM interval centered around bz, 5 are nonbreakers (Figure 5 ) . A possible interpretation of this observation is that the two elements in the nonbreaking pairs are in reverse orientation relative to each other. In such a configuration, the outsideends of the composite transposon structure would be identical and would notberecognized by the transposase, since an Ac element with two identical endscannot transpose et al. 1989). The apparent excess ofdirect (COUPLAND over inverted short-range transpositions of Ac from bz-m2(Ac), acorollary of the aboveinterpretation, would remain unaccounted for at this time. Alternatively, if in a composite transposon structure with inverted terminal TEs, one outside end and one inside end in each of the TEs can form a functional transposition complex with the transposase, dicentric and acentric chromatids could also be generated by events analogous to those shown in Figure 6. Based on this consideration, a second possible interpretation of thedata is that,amongthosetransposon pairs analyzed, all the closely linked ones are breakers and
H. K. Dooner and A. Belachew
860
I
AD
EB
H NE
the occurrence of non-breaking pairs in the vicinity of bz (Table 1 ; Figure 5 ) is due to factors other than the relative orientation of theelements. An examination of the datain Table 1 reveals that the distribution of breakersandnonbreakers is not symmetrical around the bz locus. Of the 17 trAcs that lie in the interval -1.1 to +3.3 cM and behave as breakers in conjunction with either Ac or Ds at bz, 1 1 are located in the continuous segment -1.1 to +0.3 cM. It could it is notthe bearguedthat beyondthisinterval, element at bz, but other Ac ends scattered in 9s that form breaking structures with Acs that have transposed nearby. Depending on the accuracy of the estimates of genetic distances, at least 2 such Ac ends would have to be postulated in the interval between +0.4 and +3.3 cM to account for the discontinuities in the distribution of breaking trAcs seen in that interval. T h e -1.1 to +0.3 cM interval within which a trAc would form a breaking pairwith an element at bz is asymmetrical about 6%:the T E a tbz would form breaking pairswith trAcs located both proximally and distally, but the genetic distance separating the two elements would be 3-4 times higher on thedistal than the proximal side of the gene. It is possible that the maximum physical length separating the ends of a MTn is the same on eitherside of bz and that genetic lengths onthe distal side of bz represent smaller physical lengths than on the proximal side. We have found, in fact, that the amountof recombination per kb is higher in the distal part of the bz gene, which is flanked by several kb of single- or low-copy DNA, than in the proximal sideof the gene,which is flanked by repetitive DNA (H. K. DOONER,D. BURG= and E. RALSTON, unpublished observations). There is also precedent for the existence of Ac ends scatteredin the genome since many sequences that share homology with theends of Ac have been detected inmaize (FEDOROFF, WESSLER and SHURE1983). It should be possible to determine the orientation of the two closely linked elements around bz by an
CF
" Dl
:+:::fi.y,..: . .:... . .......... ....
D l . . . ..
.......... ......
CF
....... H M.:............. .HDI
AD
EC
................
BF
............ ...... ...........Kl M H-, H .
AD
EC
. .: . . .
...... ........ ........ M H.............
EA'
H
0
FIGURE6.-Model for chromosome breakage based on the transposition of a partially replicated macrotransposon (MTn). a, A replicon in a replicating chromosome. The centromere is the open circle on the left. The body of the partially replicated MTn is shaded and the left and right termini of the directly repeated transposable elements at either end of the MTn are indicated with solid and open arrow-heads, respectively, to denote orientation. The partially replicated MTn B-C excises and reinserts at site EF. T h e transposase cuts at the endsof the MTn sites AB and CD, and at the target insertion site EF. The 5' ends of each cut are marked as dots to indicate strand polarity. b, Products of the reinsertion of the MTn in direct orientation relative to the original insertion: a chromosome with a transposed MTn and a chromosome with a deletion of the segment D-E adjacent to the MTn. c, Products of the reinsertion of the MTn in inverse orientation relative to the original insertion: a chromosome with a transposed inverted MTn and a dicentric chromosome. The dicentric will break during cell division and set up a chromatid breakage-fusion-bridge cycle.
Breakage Chromosome examination of unusual excision products that result in loss of the mutable phenotype. Thus, transposons in direct orientation should produce interstitial deletions of the segment between, and including, the two transposons (Figure 6a; RALSTON,ENGLISHand DOONER1989), whereas transposons in reverse orientation should produce either inversions that retain both transposons or deletions that retain only one. In practice, however, it is often difficult to recover large deletions in plants because they are selected against during the haploid gametophytic phase of the plant’s life cycle. The existence of viable deletions that include the segment between bz and the distal marker sh (MCCLINTOCK 1956a; MOTTINCER1973)argues that it should be possible to recover deletions between bz and trAc elements located about 2-3 cM distal to bz. Another factor to consider in seeking an explanation for the failure of some pairs of closely linked TEs to act as breakersis the relative timing of replication of their DNA during the S phase of the cell cycle. There is genetic and molecular evidence that transposition is associated with DNA replication (GREENBLATT and BRINK 1962; CHEN, GREENBLATT and DELLAPORTA 1988; LAUFSet a/. 1990). If the two closely linked TEs belonged to replicon families that replicated their DNA at different times in S phase (VAN’T HOF1988), their endswould not serve as cosubstrates for the Ac transposase and the two TEs would not act as a macrotransposon. In our study, however, all the trAcs that were tested for breakage in combination with a second element at bz came originally from bz. If competent trAc receptor sites are those that undergo replication at thesame time in S phase as the Ac donor site, it follows that the ends of the element at bz and of the closely linked trAcs replicate synchronously and can function potentially as the ends of a macrotransposon. Regardless of the mechanism of chromosome breakage, the results reported here can explain the relative stabilities of different Ds elements. MCCLINTOCK (1947,1948,1949) describedstateIelementsthat produce many breaks and few reversions when inserted in a locus and state I1 elements that produce few, if any, breaks and frequent reversions. State 11 is very stable, whereas state I is unstable and changes frequently to state 11. The single Ds and Ac elements in bz-mZ(DZ) and bz-mZ(Ac), respectively, behave as state I1 elements, but the cis combination of either element with a closely linked trAc behaves as a state I composite. Similarly, the cis combination of two Ds elements should also behave as a state I composite in the presence of Ac. One such chromosome-breaking, compound Ds has been described in the sh-m5933 mutable allele (DORINGet al. 1989). A stateI composite would be unstable since loss of one of the elements
by Transposons
86 1
by either transposition or crossing over would convert it intoastable,non-breakingstate I1 element. In Drosophila, where chromosomal rearrangements are associated with the activity of P and FB elements (BINGHAM and ZACHAR 1989; ENGELS1989), and in Antirrhinum, where they are associated with Tam3 activity (MARTIN,MACKAYand CARPENTER1988; ROBBINS,CARPENTER and COEN 1989), composite structures such as those described here may also play a role in producing chromosome breaks. Since the TEs Ac and Spm/En (Suppressor-mutator/ Enhancer) tend to transpose preferentially to nearby sites (MCCLINTOCK 1956a,b; VAN SCHAIKand BRINK 1959); GREENBLATT 1984; DOONERand BELACHEW 1989; NOVICKand PETERSON 1981 ;JONES et al. 1990; DOONER et al. 199 l),chromosome breaking structures consisting of closely linked transposon pairs must be generated often when Ac or Spm/En transpose. Thus, a negative selection pressure will be imposed upon this class of transposition product, which will lead to a decrease in the rate with which transposons spread in thegenome,but atthe same time, to a wider dispersal of the elements throughout thegenome. We thank JOYCE HAYASHI for the artwork and DIANEBURGESS, RICHJORGENSEN, JANIS KELLER, ED RALSTON,TIMROBBINS and GARY WARREN for comments on the manuscript. This is paper No. 5-1 7from DNA Plant Technology.
LITERATURE CITED BINGHAM, P., and Z. ZACHER,1989 Retrotransposons and the FB transposon from Drosophila melanogaster, pp. 485-502 in Mobile DNA, edited by D. BERGand M. HOWE.American Society for Microbiology, Washington, D.C. BRINK,R. A., and R. A. NILAN, 1952 The relation between light variegated and medium variegated pericarp in maize. Genetics 37: 5 19-544.
CHEN, J.,
I.
M.
GREENBLATT and
S.
L.
DELLAPORTA,
1988 Transposition of Ac from the P locus of maize into unreplicated chromosomal sites. Genetics 117: 109-1 16.
COUPLAND, G., C. PLUM,S. CHATTERJEE, A. POSTand P. STARLINGER, 1989 Sequences near the termini are required for transposition of the maize transposon Ac in transgenic tobacco plants. Proc. Natl. Acad. Sci. USA 86: 9385-9388. DOONER, H. K., and A. BELACHEW, 1989 Transposition pattern of the maize element Ac from the bz-m2(Ac) allele. Genetics 122: 447-457.
DOONER, H. K., J. ENGLISH and E. J. RALSTON, 1988 The frequency of transposition of the maize element Activator is not affected by an adjacent deletion. Mol. Gen. Genet. 211: 485491.
DOONER,H. K., E. J. RALSTON and J. ENGLISH,1988 Deletions and breaks involving the borders of the Ac element in the brm2(Ac) allele of maize, pp. 2 13-226 in International Symposium on Plant Transposable Elements, edited by 0.E. NELSONand C. WILSON.Plenum Press, New York. DOONER, H. K., J. ENGLISH, E.J. RALSTON and E. WECK,1986 A single genetic unit specifies two transposition functions in the maize element Activator. Science 2 3 4 2 10-2 1 1. DOONER,H. K., J. KELLER, E. HARPERand E. RALSTON, 199 1 Variable patterns of transposition of the maize element Activator in tobacco. Plant Cell 3: 473-482.
862
H. K. Dooner and A. Belachew
DORING,H.-P., B. NELSEN-SALZ, R. GARBERand E. TILLMAN, 1989 Double Ds elements are involvedinspecific chromosome breakage. Mol. Gen. Genet. 219: 299-305. ENGELS, W.,1989 P elements in Drosophila melanogaster, pp. 437484 in Mobile DNA, edited by D. BERGand M. HOWE.American Society for Microbiology, Washington, D.C. FEWROFF,N., S. WESSLERand M. SHURE,1983 lsolation of the transposable controlling elements Ac and Ds. Cell 3 5 235-242. GREENBLATT, I. M., 1984 A chromosome replication pattern deduced from pericarp phenotypes resulting from movements of the transposable element Modulator inmaize. Genetics 108: 471-485. GREENBLATT, I. M., and R. A. BRINK,1962 Twin mutations in medium variegated pericarp maize. Genetics 47: 489-50 1. JONES, J., F. CARLAND, E. LIM, E.RALSTON and H. K. DOONER, 1990 Preferential transposition of the maize element Activator to linked chromosomal locations in tobacco. Plant Cell 2: 701707. LAUFS, J., U. WIRTZ,M. KAMMANN, V. MATZEIT, S. SHAEFER, J. SCHELL, A. P. CZERNILOFSKY, B. BAKER and B. GRONENBORN, 1990 Wheat dwarf virus Ac/Ds vectors: expression and excision of transposable elements introduced into various cereals by a viral replicon. Proc. Natl. Acad. Sci. USA 87: 7752-7756. MARTIN,C.,S. MACKAYand R. CARPENTER, 1988 Large scale chromosomal restructuring is induced by the transposable element Tam3 at the nivea locus of Antirrhinum majus. Genetics 119 171-184. MCCLINTOCK, B., 1947 Cytogenetic studies of maize and Neurospora. Carnegie Inst. Washington Year Book 46: 146-152. MCCLINTOCK, B., 1948 Mutable lociin maize. Carnegie Inst. Washington Year Book 47: 155-169. MCCLINTOCK, B., 1949 Mutable lociinmaize. Carnegie Inst. Washington Year Book 48: 142-154. MCCLINTOCK, B., 1951 Chromosome organization and gene expression. Cold Spring Harbor Symp. Quant. Biol. 16: 1347. MCCLINTOCK, B., 1955 Controlled mutation inmaize. Carnegie Inst. Washington Year Book 55: 323-332.
MCCLINTOCK, B., 1956a Mutation in maize. Carnegie Inst. Washington Year Book 5 5 323-332. MCCLINTOCK, B., 1956b Controlling elements and the gene. Cold Spring Harbor Symp. Quant. Biol. 21: 197-216. MCCLINTOCK, B., 1962 Topographical relations between elements of control systems in maize. Carnegie Inst. Washington Year Book 61: 448-461. MOTTINGER, J., 1973 Unstable mutants of bronze induced by premeiotic X-ray treatment in maize. Theor. Appl. Genet. 43: 190- 195. MURRAY,J. A,, 1987 Bending the rules: the 2 p plasmid of yeast. Mol. Microbiol. 1: 1-4. NOVICK,E., and P. A. PETERSON,1981 Transposition of the Enhancer controlling element system in maize. Mol. Gen. Genet. 183: 440-448. RALSTON, E. J., J. ENGLISH and H. K. DOONER,1987 Stability of deletion, insertion and point mutations at the bronze locus in maize. Theor. Appl. Genet. 7 4 471-475. RALSTON, E. J., J.ENGLISHand H. K. DOONER, 1988 Sequence of three bronze alleles of maize and correlation with the genetic fine structure. Genetics 119: 185-197. RALSTON, E. J., J. ENGLISH and H. K. DOONER, 1989 Chromosome-breaking structure inmaizeinvolving a fractured Ac element. Proc. Natl. Acad. Sci. USA 86: 94519455. RHOADES, M. M., 1952 The effect of the bronze locus on anthocyanin formation in maize. Am. Nat. 8 6 105-108. ROBBINS, T . , R. CARPENTER and E. COEN,1989 A chromosome rearrangement suggests that donor and recipient sites are associated during Tam3 transposition in Antirrhinummajus. EMBO J. 8: 5-13. VAN SCHAIK, N., and R. A. BRINK,1959 Transposition of Modulator, a componentof the variegated pericarp in maize.Genetics 44: 725-738. VAN'THOF,J., 1988 Chromosomal replicons of higher plants, pp. 75-89 in Chromosome Structure and Function, edited by J. P. GUSTAFSON and R. APPELS. Plenum Press, New York. Communicating editor: W. F. SHERIDAN
Chromosome Breakage byPairs of Closely Linked Transposable Elements of the Ac-Ds Family in Maize Hugo K. Dooner and Alemu Belachew DNA Plant Technology Corporation, Oakland, Calqornia 94608-1239 Manuscript received November 20, 1990 Accepted for publication August 2, 1991
ABSTRACT Chromosome breaks and hence chromosomal rearrangements often occur in maize stocks harboring transposable elements (TEs), yet it is not clear what types of TE structures promote breakage. We have shown previouslythat chromosomes containing a complex transposon structure consisting of an Ac (Activator) element closely linked in direct orientation to aterminally deleted or fractured Ac CfAc) element have a strong tendency to break during endosperm development. Here we show that pairs of closely linked transposons with intact ends, either two Ac elements-a common product of Ac transposition-or an Ac and a Ds (Dissociation) element, can constitute chromosome-breaking structures, and that the frequency of breakage is inversely related to intertransposon distance. Similar structures may also be implicated in chromosome breaks in other eukaryotic TE systems known to produce chromosomal rearrangements. The present findings are discussed in light of a model of chromosome breakage that is based on the transposition of a partially replicated macrotransposon delimited by the outside ends of the two linked TEs.
T
HOUGH transposable elements (TEs) in maize werefirst detected by MCCLINTOCK (1947, 1948, 1949) because of their ability to break chromosomes, it is only recently that particular TE structures have been implicated in chromosome breakage (RALSTON,ENGLISHand DOONER 1989; DORINGet al. 1989). A common feature of these structures is that they are compound, ie., they contain more than one element and,therefore, multiple T E ends inclose proximity. One structure, containing three TE ends, consists of two closely linked, but recombinationally separable, elements in direct orientation: a 4.6-kilobase (kb) intact Activator (Ac) element and a 2.5-kb fractured Ac (JAc) element that is deleted for the end ENGLISH that lies closest to the intact Ac (RALSTON, and DOONER 1989). Another one,containing four T E ends, consists of two identical copies aof2-kb internal deletion derivative of Ac, arranged so that onecopy is inserted in inverse orientation into the second copy (DORINGet al. 1989). The former structure arose in a bz-m2(Ac) stock. The bz-m2(Ac) allelehas an autonomous (ie., selfmobile) Ac element inserted in the second exon of the bz gene (RALSTON, ENGLISH and DOONER1988) and can mutate t o stable bz alleles as a consequence of Ac transposition (MCCLINTOCK 1956a,b; DOONERand BELACHEW 1989). One such allele, bz-s:2094, carries a 2.5-kb fractured Ac CfAc) element at the previous site of insertion of Ac and, in addition, a transposed Ac(trAc) 0.05 centimorgans (cM) proximal to fAc (DOONER, RALSTONand ENGLISH1988). In lines carrying the bz-s:2094 allele, chromosome Genetics 129: 855-862 (November, 1991)
breaks at or near bz occur with high frequency during the mitotic divisions that give rise to the endosperm. ENGLISHand DOONER We have proposed (RALSTON, 1989) that chromosome breakage is due to the transpositionof a partially replicated macrotransposon (MTn) extending from the distal end of fAc to the proximal end of the trAc (Figure 1) based on the following observations: (1) Breaks occur only if fAc and Ac are closely linked. (2) A deletion carrying an “empty” MTn excision site was recovered and characterized by sequencing the new junction. (3) The fAc and Ac elements are in direct orientation, as wouldbe expected if fAc and Ac provided the left and right termini of the MTn. The key feature of our model is not the singleended fAc element at one end of the MTn, but the presence of opposite left and right transposon ends separated by a stretch of chromosome (the body of the transposon) that is in a different replication state from one or both transposon ends atthe time of transposition. The model predicts that the combination of two homologous transposons that are closely linked and in direct orientation (thus providing oppositeleft and right ends) will causechromosome breaks. Here, we have examined the chromosomebreaking properties of paired transposons, one at a constant location in the bronze ( b z ) locus and the second one at variable genetic distances from bz. MATERIALS AND METHODS Geneticstocksandcrosses: All the stocks used in the present investigation have the genetic background of the
856
fAc
9s
H. K. Dooner and A. Belachew L
e"rn
TABLE 1
R 1
-
I
"
wx
List of derivatives carrying a trAe
(0.05 cM)
FIGURE 1.--Diagram of thecompositetransposonstructure found in the short arm of chronlosome 9 of the bz-s:20Y4 allele. This structure, which can be excised as ;I Ill;1crotrallsposoIl (hlTn) rstentling from the left (L) end of a fAr element to the right ( K ) end of an intact Ar element located 0.05 cM prositnallv, has been post ul;ltect to promote cI1ronwsonIe breaks.
inbred W22 and carry one of the following allelesat the bz locus in the short arm of chromosome 9 (9.9. Bz-McC (purple): the normal progenitor allele of the T E mutations used in this study. bz-mZ(Ac) (bronze-purple variegation): an allele arising from insertion of a 4.6-kb Ac element in the second exon of Bz-McC (MCCLINTOCK1955; RALSTON, ENGLISH and DOONER1988). bz-mZ(DI) (bronze in the absence of Ac; spotted in its presence): the first derivative from bz-mP(Ac); it harbors a 3.3-kb Ds element as a consequence of an internal deletion in Ac (MCCLINTOCK 1962; DOONERet al. 1986). bz-R (bronze): the stable, reference allele at the locus; it is associated with a 340 bp deletion in the transcribed region (RHOADES1952; RALSTON, ENGLISH and DOONER1987). bz-s:2094(fAc) Ac2094 (bronze): a stable derivative of bzm2(Ac) harboring a 2.5-kb terminally deleted or fractured Ac (fAc) element at bz and a trAc at a location 0.05 CM proximal to bz (DOONER,RALSTONand ENGLISH1988). bz-s:2094(fAc) (bronze): a derivative of bz-s:2094(fAc) Ac2094 that has lost the trAc2094 element (RALSTON,ENGLISH and DOONER1989). bz-s:2114(Ac) (bronze): a stable derivative of bz-mZ(Ac) that arose by deletion of 789 bp of bz sequence proximally and RALSTON1988). adjacent to Ac (DOONER, ENGLISH Two classes of chromosomes derived from the above stocks were tested for breakage: (1) bz-mZ(Ac) WAC,chromosomes with two linked Acs, arose directly by transposition of Ac in a bz-mZ(Ac) stock. The increased Ac dosage results in a finely spotted phenotype. (2) bz-mZ(DI) trAc, chromosomes with a Ds element and a linked WAC,arose by recombination between the bz-mZ(DI) reporter allele and a trAc element in experiments designed to map the location of different trAcs relative to bz. To monitor chromosome breaks genetically, the above stocks (all C in constitution) were used to pollinate c Rz ear parents. Kernels where no breaks occur will be purple; those where breaks occur in 9s will display colorless sectors in a purple background (see RESULTS). The kernels from the various testcrosses were examined individually under 7 X magnification with a stereo-zoom microscope, which allowed the resolution of very small colorless sectors (4-8 cells and larger). In general, the size of the colorless sectors seen in kernels carrying two Ac elements was smaller than in kernels carrying one Ac and one Ds element, an observation in agreement with the well documented negative dosage effect of Ac in maize. An increase in Ac dosage results in a delay of the timing and a reduction of the frequency of transposition (MCCLINTOCK 195 1; BRINKand NILAN1952). In each cross, at least 500 kernels were scored for sectors.
RESULTS
Chromosomes having pairs of homologous elements, one at a fixed location in the bz gene and the other one at variable distances away from bz, were
61 I0 7084 6084 2107 6085 7066 61 17 2097 2 106 7080 3137 7068 7067 7072 707 1 7065 7056 21 16 61 14 2094 2094 6087 6058 6067 7082 2103 7074 705 1 6083 708 1 7070 7069 Controls Ac
Ds2 ( D I ) 21 14 fAc
bz-m2 ( A c ) bz-m2 ( A r ) bz-m2 ( D l ) bz-m2 (Dl) bz-m2 ( D I ) bz-m2 ( A r ) bz-m2 ( A r ) bz-m2 (Dl) bz-m2 ( D I ) bz-m2 ( A c ) bz-m2 ( D I ) bz-m2 ( A c ) bz-m2 ( A c ) bz-m2 ( A c ) bz-m2 ( A r ) bz-m2 ( A r ) bz-m2 ( D l ) bz-m2 ( D I ) bz-m2 ( A r ) bz-m2 (Dl) b~-S:2O94( f A c ) bz-m2 ( D I ) bz-m2 ( D I ) bz-m2 ( D I ) bz-m2 ( A r ) bz-m2 ( D I ) bz-m2 ( A c ) bz-m2 ( D I ) bz-m2 ( D I ) bz-m2 ( A c ) bz-m2 ( A c ) bz-m2 ( A c )
+30.8 +27.7
1.4 f 0..5 1.1 f 1 . 1
+23.5 +21.2 +17.5 + I 2.0 +4.7 +3.3 +3. 1 +2.1 + I .4 + I .3 +1.2 + I .2 + I .0
1.5 f 0.6 7.6 f 1.4 1.1 f 1.1 1 .x f 0.7 1.4 f 0.8 24 f 2.0 26 f 2.5 16 f 2.0 17 f 1.3 32 f 2.1 3.3 f 0.8 4.3 f 0.9 38 f 2.7 2.0 f 0.5 2.0 f 0.7 5 1 f 2.3 82 f 1.8 86 f 1 . 1 92 f 1.0 67 f 3.1 83 f 3.2 84 f 3.6 46 f 3.1 87 f 2.3 18 f 1.4 71 f 1.7 56 f 3.2 4 f 2 5.7 f 1.3 1 .0 f 0.7
bz-m2 (Ac) bz-m2 ( D I ) b 1 - ~ : 2 1 1 4( A c ) bz-s:2094 (fAC)
0.0
+0.6 +0.4
+0.3 +0.2 +0.05 +0.05
-0. I -0.2 -0.3 -0.3 -0.5 -0.5 -0.8 -1.1
-2.2 -4.1 -16.9
0.0
5 . 6 f 0.6 0.07 f 0.07 5.3 f 1.2 0.1 a 0.07
a Locations proximalto bz are represented with positive numbers and those distal to bz with negative numbers. * Proportion of kernels with >5 c sectors.
recovered in the course of a study that examined the pattern of transposition of Ac from bz-m2(Ac) to linked sites (DOONERand BELACHEW1989). They fall into two genetic classes: (i) bz-m2(Ac) trAc, chromosomes having two linked Ac elements, and (ii) bz-m2(DI) trAc, chromosomes having a nonautonomous Ds (Dissociation) element at 6 2 and a linked trAc. The 3.3-kb Ds element in bz-m2(DI) is an internal deletion derivative of Ac in the same location and orientation as the Ac element in bz-m2(Ac) (DOONERet al. 1986). The chromosomes with two linked transposons that were tested for breakage are listed in Table 1. T h e y carry a trAc and are identified by the isolation number of the trAc stock. Their bz locus constitution andthe distance between bz and the trAc (DOONERand BELACHEW 1989) are also given in Table 1. We know that in all these chromosomes both ele-
Breakage Chromosome C
C
82
C
h?
Bz
-0
FIGURE2,“Genetic cross used to monitor the loss of the dominant marker C (colored aleurone) located distal to bz in the short arm of chromosome 9 . A c tester is pollinated with C stocks carrying one transposon at bz and anotherone at variable distances from bz. The frequency of colorless sectors in thealeurone provides an estimate of the frequency of breaks that occur during development of the aleurone.
by Transposons
857
linked Acs. All kernels fromcrosses monitoring breakage were scored as having either > o r 5 5 c sectors and the results were expressed in terms of p , the proportion of kernels with >5 c sectors (Table 1).The standard error of p was computed by the formula JP(1 - P)/n. T h e results of the tests for breakage are summarized in Figure 4, where the “efficiency of breakage” of the various chromosomes being tested, expressed as percent of kernels with >5 c sectors, is plotted vs. the genetic distance separating the two transposons, expressed in cM. It is evident that only transposon pairs that are closely linked are capable of producing more chromosome breaks than the single Ac control. Furthermore, among the “breakers” there is an inverse relationship between breakage frequency and inter-transposon distance, so that the higher breakage frequencies (>50%) are given only by combinations of elements separated by 1 cM or less. This can be more clearly seen in Figure 5, which is a close-up of the 5-cM region on either side of bz. Transposed Acs that, in combination with a second element atbz, cause significantly more breaks than the bz-m2(Ac) control are connected with a line in Figure 5 . They map within 3.3 cM from bz. Chromosomes of a bz-m2(DZ) trAc type appeartoproducemore c sectors than chromosomes of a bz-m2(Ac) trAc type. This is to be expected given the negative dosage effect of Ac on transposition frequency observed in maize (McCLINTOCK1951; BRINKand NILAN 1952). Interestingly, not all closely linked transposon pairs behave as chromosome breakers. Of 22 trAcs tested in the interval that lies 3.3 cM on either side of bz, 5 are clearly nonbreakers (Table 1). Assuming that the estimates of genetic distance are reasonably accurate (they are based on populations of >lo00 gametes; DOONER and BELACHEW 1989), these results indicate that a short inter-transposon distance between the transposon at bz and a trAc is nota sufficient conditionfor the transposon pair to cause chromosome breaks.
ments can transpose, which means that the ends of the elements are intact, in contrast to the terminally deleted fAc element of the bz-s:2094 allele. Evidence that the element atbz (either Ac or Ds) can transpose is provided simply by the bronze-purple variegation seen in somatic tissues of plants carrying these chromosomes, as well as by the germinalpurple ( B z ’ ) derivatives that occur regularly among the progenies of such plants. Evidence that the trAc can still transpose, i.e., thatsecondary transpositions of the trAc element occur, was provided by DOONERand BELACHEW (1989). The average frequency of secondary transposition of Ac from its new location in 9s to unlinked sites in the genomewas estimated to be about 5 x 10-3, which is similar to thatof Ac from its original location in the bz-m2(Ac) allele, and low enough not to interfere with the mapping of trAcs. Chromosomebreakswere assayed genetically by monitoring the loss of the dominant marker C (colored kernel, epistatic to b z ) located distal to bz in the same chromosome arm (Figure 2). Normally, kernels from a cross between a c Bz female parent and a C bz male parent will be solid purple. If the C bz chromosome containing the pair of transposons being tested breaks at or around the bz locus during endosperm DISCUSSION development, the distal marker C will be lost resulting in a colorless sector in the kernel. Thus, thefrequency We have tested herethe chromosome-breaking of colorless sectors provides an estimate of the inciproperties of pairs of closely linked transposons bedence of chromosome breaks during the cell divisions longing to the Ac-Ds family of maize TEs. One memthat produce a mature kernel. In crosses with the bzber of the pair, either Ac or Ds, occupied a constant s : 2 0 9 4 allele, which carries fAc and a trAc 0.05 cM location in the bz locus, and the second one, a trAc, apart, almost all the kernels have many (>lo) colorless was inserted in sites located at variable genetic dis(c) sectors of variable sizes, whereas in control crosses tances proximally and distally from bz. We found that with either bz-m2(Ac) or bz-m2(DZ), carrying single in many lines carrying pairsof closely linked transpotransposons, kernels with >5 c sectors are either rare sons, chromosome breaks resulting in the loss of the o r nonexistent (Figure 3). We can explain the presdistal marker C occurred during the development of ence of occasional sectors in bz-m2(Ac), and not in bzthe endosperm and that the frequency of breakage m2(DZ), kernels as being due to the generation by Ac was inversely related to the distance separating the transposition during endosperm development ofchro- t w o transposons. These results support our model of mosome-breaking structures consisting of two closely chromosome breakage (RALSTON,ENGLISH and
858
H. K. Dooner andA. Belachew to/
" "
-t- br-~:2094 ( f k ) Ac2094
bm2(Ac) bmn2(DJJ
0.9
FIGURE3.-Distribution of kernels with colorless sectors incrossesof c testers by C stocks carrying various transposable elements in 9s. bz-n2(Ac) and bzm2(DI) carry single elements, Ac and Ds,respectively, whereas bz-s:2094uAc) Ac2094carry the elementsfAc and Ac, 0.05 cM apart.
No. c sectorwkernel I
I
eo-
60-
FIGURE4,"Relationship between intertransposon distance and chromosome breaks. Chromosome breakage was assayed genetically as shown in Figure 2. The frequency of kernels with >5 c (colorless) sectors provides an estimate of breakage frequency. Locations distal to br are represented with negative numbers and those proximal to br with positive numbers (see Table 1). Stocks carrying a trAc and a Ds at b t are represented with open circles. Stocks carrying atrAc and an Ac at br are represened with solid circles.
40-
wx
cM Distal
cM Proximal
DOONER 1989) that explains breaks as a consequence of the transposition of a partially replicated MTn extending between the outside ends of two closely linked TEs. We previously considered the consequences of transposition of a MTn in which both ends, locatedin different replicons, have begun to replicate, whereas the centerhas not (Figure6 in RALSTON, ENGLISH and DOONER 1989). Both dicentric and acentric chromatids can arise as the consequences of reinsertion of the MTn in inverted orientation relative to the progenitor chromosome. Similar outcomes are also possible if bothends of theMTnare in the same replicon,
separated by a minimum distance, but only a single end has replicated when transposition occurs. We illustrate these outcomes in Figure 6, which is simply an extension of our earlier diagram that depicted a MTn with both ends replicated. Arguments for the requirement of a minimum distance separating the ends of the MTn have already been presented (RALSTON, ENGLISH and DOONER1989). Figure 6a shows a MTn (B-C) in which only one end (B) has replicated at the time of transposition. As in any other transposition, the Ac transposase cuts at three sites: the two MTn ends (AB and CD) and the target o r receptor site (EF). In Figure 6 the receptor
Chromosome Breakageby Transposons
859
FIGURE5.-Close-up of the region immediately adjacent to bz in Figure 4. Distances distal to bz are given negative numbers and those proximal, positive numbers. The lines connect those trAcs that, in combination with an element at br, give significantly more breaks than the single Ac control (bz-m2).
0
0
I
-5
-4
0
”
1
I
I
-3
-2
-1
sh
cM Distal
0
0.
.r
I
I
1
I
P
1
2
3
4
R
bi
cM Proximal
site is shown distal to the MTn, but it can also be located proximal to it. The MTn can reinsert at the EF receptor site in either direct or inverted orientation relative to its original orientation in the chromosome. If the MTn inserts in direct orientation (Figure 6b), the three new junctions are: AD, corresponding to the “empty”site, and EB and CF, corresponding to the new MTn borders. This type of transposition can be most readily visualized as a displacement of the unreplicated D-E fragment, which separates the donor and receptor sites, to the already replicated proximal end of the MTn. Following resolution of the replication forks, the two daughter chromatids carry, respectively, a distally and directly transposed MTn (i) and a deletion of the D-E fragment that was originally adjacent to thedistal end of the MTn (ii). If the MTn inserts in inverted orientation (Figure 6c),thethree new junctions are AD, again corresponding to the “empty” site, and EC and BF, corresponding tothe new MTnborders.This type of transposition sets up chasing replication forks, which have also been postulated in the mechanism of amplification of the 2u plasmid of yeast (MURRAY1987). Resolution of the replication forks would result in a chromatid with a distally transposed MTn (iii) and a dicentric chromatid (iv). However, it is likely that the trailing replication fork, which will not converge with another replication fork, is not resolved during the course of the S phase, leading to the formationof an incompletely duplicated dicentric. If B-C transposed proximally in aninvertedorientation, anacentric chromatid would be formed (not shown). The latter outcome can be visualized by simply shifting the po-
sition of the centromere shown in Figure 6 from the left to the right end of the chromosome, next to the EF receptor site. Chromatid iv in Figure 6c would be an acentricsince it lacks the D-E-F segment thatwould now be adjacent to the centromere.A dicentric chromatid will form an anaphase bridge that will break and initiate a chromatid breakage-fusion-bridge cycle and anacentric will lag at the metaphase plate and be lost from either daughter cell. Though the majority of T E pairs comprising one element at br and a trAc in the vicinity of bz are breakers, some are not. Of 22 trAcs tested in a 6-cM interval centered around bz, 5 are nonbreakers (Figure 5 ) . A possible interpretation of this observation is that the two elements in the nonbreaking pairs are in reverse orientation relative to each other. In such a configuration, the outsideends of the composite transposon structure would be identical and would notberecognized by the transposase, since an Ac element with two identical endscannot transpose et al. 1989). The apparent excess ofdirect (COUPLAND over inverted short-range transpositions of Ac from bz-m2(Ac), acorollary of the aboveinterpretation, would remain unaccounted for at this time. Alternatively, if in a composite transposon structure with inverted terminal TEs, one outside end and one inside end in each of the TEs can form a functional transposition complex with the transposase, dicentric and acentric chromatids could also be generated by events analogous to those shown in Figure 6. Based on this consideration, a second possible interpretation of thedata is that,amongthosetransposon pairs analyzed, all the closely linked ones are breakers and
H. K. Dooner and A. Belachew
860
I
AD
EB
H NE
the occurrence of non-breaking pairs in the vicinity of bz (Table 1 ; Figure 5 ) is due to factors other than the relative orientation of theelements. An examination of the datain Table 1 reveals that the distribution of breakersandnonbreakers is not symmetrical around the bz locus. Of the 17 trAcs that lie in the interval -1.1 to +3.3 cM and behave as breakers in conjunction with either Ac or Ds at bz, 1 1 are located in the continuous segment -1.1 to +0.3 cM. It could it is notthe bearguedthat beyondthisinterval, element at bz, but other Ac ends scattered in 9s that form breaking structures with Acs that have transposed nearby. Depending on the accuracy of the estimates of genetic distances, at least 2 such Ac ends would have to be postulated in the interval between +0.4 and +3.3 cM to account for the discontinuities in the distribution of breaking trAcs seen in that interval. T h e -1.1 to +0.3 cM interval within which a trAc would form a breaking pairwith an element at bz is asymmetrical about 6%:the T E a tbz would form breaking pairswith trAcs located both proximally and distally, but the genetic distance separating the two elements would be 3-4 times higher on thedistal than the proximal side of the gene. It is possible that the maximum physical length separating the ends of a MTn is the same on eitherside of bz and that genetic lengths onthe distal side of bz represent smaller physical lengths than on the proximal side. We have found, in fact, that the amountof recombination per kb is higher in the distal part of the bz gene, which is flanked by several kb of single- or low-copy DNA, than in the proximal sideof the gene,which is flanked by repetitive DNA (H. K. DOONER,D. BURG= and E. RALSTON, unpublished observations). There is also precedent for the existence of Ac ends scatteredin the genome since many sequences that share homology with theends of Ac have been detected inmaize (FEDOROFF, WESSLER and SHURE1983). It should be possible to determine the orientation of the two closely linked elements around bz by an
CF
" Dl
:+:::fi.y,..: . .:... . .......... ....
D l . . . ..
.......... ......
CF
....... H M.:............. .HDI
AD
EC
................
BF
............ ...... ...........Kl M H-, H .
AD
EC
. .: . . .
...... ........ ........ M H.............
EA'
H
0
FIGURE6.-Model for chromosome breakage based on the transposition of a partially replicated macrotransposon (MTn). a, A replicon in a replicating chromosome. The centromere is the open circle on the left. The body of the partially replicated MTn is shaded and the left and right termini of the directly repeated transposable elements at either end of the MTn are indicated with solid and open arrow-heads, respectively, to denote orientation. The partially replicated MTn B-C excises and reinserts at site EF. T h e transposase cuts at the endsof the MTn sites AB and CD, and at the target insertion site EF. The 5' ends of each cut are marked as dots to indicate strand polarity. b, Products of the reinsertion of the MTn in direct orientation relative to the original insertion: a chromosome with a transposed MTn and a chromosome with a deletion of the segment D-E adjacent to the MTn. c, Products of the reinsertion of the MTn in inverse orientation relative to the original insertion: a chromosome with a transposed inverted MTn and a dicentric chromosome. The dicentric will break during cell division and set up a chromatid breakage-fusion-bridge cycle.
Breakage Chromosome examination of unusual excision products that result in loss of the mutable phenotype. Thus, transposons in direct orientation should produce interstitial deletions of the segment between, and including, the two transposons (Figure 6a; RALSTON,ENGLISHand DOONER1989), whereas transposons in reverse orientation should produce either inversions that retain both transposons or deletions that retain only one. In practice, however, it is often difficult to recover large deletions in plants because they are selected against during the haploid gametophytic phase of the plant’s life cycle. The existence of viable deletions that include the segment between bz and the distal marker sh (MCCLINTOCK 1956a; MOTTINCER1973)argues that it should be possible to recover deletions between bz and trAc elements located about 2-3 cM distal to bz. Another factor to consider in seeking an explanation for the failure of some pairs of closely linked TEs to act as breakersis the relative timing of replication of their DNA during the S phase of the cell cycle. There is genetic and molecular evidence that transposition is associated with DNA replication (GREENBLATT and BRINK 1962; CHEN, GREENBLATT and DELLAPORTA 1988; LAUFSet a/. 1990). If the two closely linked TEs belonged to replicon families that replicated their DNA at different times in S phase (VAN’T HOF1988), their endswould not serve as cosubstrates for the Ac transposase and the two TEs would not act as a macrotransposon. In our study, however, all the trAcs that were tested for breakage in combination with a second element at bz came originally from bz. If competent trAc receptor sites are those that undergo replication at thesame time in S phase as the Ac donor site, it follows that the ends of the element at bz and of the closely linked trAcs replicate synchronously and can function potentially as the ends of a macrotransposon. Regardless of the mechanism of chromosome breakage, the results reported here can explain the relative stabilities of different Ds elements. MCCLINTOCK (1947,1948,1949) describedstateIelementsthat produce many breaks and few reversions when inserted in a locus and state I1 elements that produce few, if any, breaks and frequent reversions. State 11 is very stable, whereas state I is unstable and changes frequently to state 11. The single Ds and Ac elements in bz-mZ(DZ) and bz-mZ(Ac), respectively, behave as state I1 elements, but the cis combination of either element with a closely linked trAc behaves as a state I composite. Similarly, the cis combination of two Ds elements should also behave as a state I composite in the presence of Ac. One such chromosome-breaking, compound Ds has been described in the sh-m5933 mutable allele (DORINGet al. 1989). A stateI composite would be unstable since loss of one of the elements
by Transposons
86 1
by either transposition or crossing over would convert it intoastable,non-breakingstate I1 element. In Drosophila, where chromosomal rearrangements are associated with the activity of P and FB elements (BINGHAM and ZACHAR 1989; ENGELS1989), and in Antirrhinum, where they are associated with Tam3 activity (MARTIN,MACKAYand CARPENTER1988; ROBBINS,CARPENTER and COEN 1989), composite structures such as those described here may also play a role in producing chromosome breaks. Since the TEs Ac and Spm/En (Suppressor-mutator/ Enhancer) tend to transpose preferentially to nearby sites (MCCLINTOCK 1956a,b; VAN SCHAIKand BRINK 1959); GREENBLATT 1984; DOONERand BELACHEW 1989; NOVICKand PETERSON 1981 ;JONES et al. 1990; DOONER et al. 199 l),chromosome breaking structures consisting of closely linked transposon pairs must be generated often when Ac or Spm/En transpose. Thus, a negative selection pressure will be imposed upon this class of transposition product, which will lead to a decrease in the rate with which transposons spread in thegenome,but atthe same time, to a wider dispersal of the elements throughout thegenome. We thank JOYCE HAYASHI for the artwork and DIANEBURGESS, RICHJORGENSEN, JANIS KELLER, ED RALSTON,TIMROBBINS and GARY WARREN for comments on the manuscript. This is paper No. 5-1 7from DNA Plant Technology.
LITERATURE CITED BINGHAM, P., and Z. ZACHER,1989 Retrotransposons and the FB transposon from Drosophila melanogaster, pp. 485-502 in Mobile DNA, edited by D. BERGand M. HOWE.American Society for Microbiology, Washington, D.C. BRINK,R. A., and R. A. NILAN, 1952 The relation between light variegated and medium variegated pericarp in maize. Genetics 37: 5 19-544.
CHEN, J.,
I.
M.
GREENBLATT and
S.
L.
DELLAPORTA,
1988 Transposition of Ac from the P locus of maize into unreplicated chromosomal sites. Genetics 117: 109-1 16.
COUPLAND, G., C. PLUM,S. CHATTERJEE, A. POSTand P. STARLINGER, 1989 Sequences near the termini are required for transposition of the maize transposon Ac in transgenic tobacco plants. Proc. Natl. Acad. Sci. USA 86: 9385-9388. DOONER, H. K., and A. BELACHEW, 1989 Transposition pattern of the maize element Ac from the bz-m2(Ac) allele. Genetics 122: 447-457.
DOONER, H. K., J. ENGLISH and E. J. RALSTON, 1988 The frequency of transposition of the maize element Activator is not affected by an adjacent deletion. Mol. Gen. Genet. 211: 485491.
DOONER,H. K., E. J. RALSTON and J. ENGLISH,1988 Deletions and breaks involving the borders of the Ac element in the brm2(Ac) allele of maize, pp. 2 13-226 in International Symposium on Plant Transposable Elements, edited by 0.E. NELSONand C. WILSON.Plenum Press, New York. DOONER, H. K., J. ENGLISH, E.J. RALSTON and E. WECK,1986 A single genetic unit specifies two transposition functions in the maize element Activator. Science 2 3 4 2 10-2 1 1. DOONER,H. K., J. KELLER, E. HARPERand E. RALSTON, 199 1 Variable patterns of transposition of the maize element Activator in tobacco. Plant Cell 3: 473-482.
862
H. K. Dooner and A. Belachew
DORING,H.-P., B. NELSEN-SALZ, R. GARBERand E. TILLMAN, 1989 Double Ds elements are involvedinspecific chromosome breakage. Mol. Gen. Genet. 219: 299-305. ENGELS, W.,1989 P elements in Drosophila melanogaster, pp. 437484 in Mobile DNA, edited by D. BERGand M. HOWE.American Society for Microbiology, Washington, D.C. FEWROFF,N., S. WESSLERand M. SHURE,1983 lsolation of the transposable controlling elements Ac and Ds. Cell 3 5 235-242. GREENBLATT, I. M., 1984 A chromosome replication pattern deduced from pericarp phenotypes resulting from movements of the transposable element Modulator inmaize. Genetics 108: 471-485. GREENBLATT, I. M., and R. A. BRINK,1962 Twin mutations in medium variegated pericarp maize. Genetics 47: 489-50 1. JONES, J., F. CARLAND, E. LIM, E.RALSTON and H. K. DOONER, 1990 Preferential transposition of the maize element Activator to linked chromosomal locations in tobacco. Plant Cell 2: 701707. LAUFS, J., U. WIRTZ,M. KAMMANN, V. MATZEIT, S. SHAEFER, J. SCHELL, A. P. CZERNILOFSKY, B. BAKER and B. GRONENBORN, 1990 Wheat dwarf virus Ac/Ds vectors: expression and excision of transposable elements introduced into various cereals by a viral replicon. Proc. Natl. Acad. Sci. USA 87: 7752-7756. MARTIN,C.,S. MACKAYand R. CARPENTER, 1988 Large scale chromosomal restructuring is induced by the transposable element Tam3 at the nivea locus of Antirrhinum majus. Genetics 119 171-184. MCCLINTOCK, B., 1947 Cytogenetic studies of maize and Neurospora. Carnegie Inst. Washington Year Book 46: 146-152. MCCLINTOCK, B., 1948 Mutable lociin maize. Carnegie Inst. Washington Year Book 47: 155-169. MCCLINTOCK, B., 1949 Mutable lociinmaize. Carnegie Inst. Washington Year Book 48: 142-154. MCCLINTOCK, B., 1951 Chromosome organization and gene expression. Cold Spring Harbor Symp. Quant. Biol. 16: 1347. MCCLINTOCK, B., 1955 Controlled mutation inmaize. Carnegie Inst. Washington Year Book 55: 323-332.
MCCLINTOCK, B., 1956a Mutation in maize. Carnegie Inst. Washington Year Book 5 5 323-332. MCCLINTOCK, B., 1956b Controlling elements and the gene. Cold Spring Harbor Symp. Quant. Biol. 21: 197-216. MCCLINTOCK, B., 1962 Topographical relations between elements of control systems in maize. Carnegie Inst. Washington Year Book 61: 448-461. MOTTINGER, J., 1973 Unstable mutants of bronze induced by premeiotic X-ray treatment in maize. Theor. Appl. Genet. 43: 190- 195. MURRAY,J. A,, 1987 Bending the rules: the 2 p plasmid of yeast. Mol. Microbiol. 1: 1-4. NOVICK,E., and P. A. PETERSON,1981 Transposition of the Enhancer controlling element system in maize. Mol. Gen. Genet. 183: 440-448. RALSTON, E. J., J. ENGLISH and H. K. DOONER,1987 Stability of deletion, insertion and point mutations at the bronze locus in maize. Theor. Appl. Genet. 7 4 471-475. RALSTON, E. J., J.ENGLISHand H. K. DOONER, 1988 Sequence of three bronze alleles of maize and correlation with the genetic fine structure. Genetics 119: 185-197. RALSTON, E. J., J. ENGLISH and H. K. DOONER, 1989 Chromosome-breaking structure inmaizeinvolving a fractured Ac element. Proc. Natl. Acad. Sci. USA 86: 94519455. RHOADES, M. M., 1952 The effect of the bronze locus on anthocyanin formation in maize. Am. Nat. 8 6 105-108. ROBBINS, T . , R. CARPENTER and E. COEN,1989 A chromosome rearrangement suggests that donor and recipient sites are associated during Tam3 transposition in Antirrhinummajus. EMBO J. 8: 5-13. VAN SCHAIK, N., and R. A. BRINK,1959 Transposition of Modulator, a componentof the variegated pericarp in maize.Genetics 44: 725-738. VAN'THOF,J., 1988 Chromosomal replicons of higher plants, pp. 75-89 in Chromosome Structure and Function, edited by J. P. GUSTAFSON and R. APPELS. Plenum Press, New York. Communicating editor: W. F. SHERIDAN
