Identification of a Regulatory Transposon That Controls Themutator Transposable Element System in Maize

Identification of a Regulatory Transposon That Controls theMutator Transposable Element System in Maize

Paul Chomet,*” Damon Lisch,* KristineJ. Hardeman: Vicki L. Chandler? and Michael Freeling* *Deflartment of Plant Biology, University of Calqornia, Berkeley, Calqornia 94720, and tInstitute of Molecular Biology, University of Oregon, Eugene, Oregon 97403 Manuscript received April 16, 1991 Accepted for publication May 22, 1991

ABSTRACT The Mutator system of maize consists of more than eight different classes of transposable elements each of which can be found in multiple copies.All Mu elements share the -220-bp terminal inverted repeats, whereas each distinct element classis defined by its unique internal sequences.The regulation of instability of this system has been difficult to elucidate due to its multigenic inheritance. Herewe present genetic experimentswhich demonstrate that thereis a single locus,MuRI, which can regulate the transposition of Mu1 elements. We describe the cloning of members of a novel class of Mu elements, MuR, and demonstrate that a member of the class is the regulator of Mutator activity, MuRI. This conclusion is based on several criteria: MuRl activity and a MuR-homologous restriction fragment cosegregate; when MuRl undergoes a duplicative transposition, an additionalMuR restriction fragmentis observed, andMuRl activity and the cosegregatingMuR fragment are simultaneously lost within clonal somatic sectors. In addition, the MuR element hybridizes to transcripts in plants with Mutator activity. Our genetic- experiments demonstrate that the MuRl transposon is necessary to specify Mutator activity in our lines.

ERTAIN lines of maize, called “Mutator lines,” a 9-bp genomic duplication upon insertion (BARKER C generate a high frequency of mutations (ROB- et al. 1984; CHENet al. 1987; SCHNABLE,PETERSON ERTSON 1978). Mutations arising in Mutator lines and SAEDLER1989). At leastsix different elements often exhibit somatically unstable phenotypes charac- representing four classes (Mul, Mul.7, Mul-del, Mu3, teristic of transposable element-induced alleles. The Mu7 and Mu8) are capable of transposition (STROM- molecular characterization of Mutator induced muta- MER et al. 1982; TAYLORand WALBOT1987; OISHI tions at Adhl, Brl, A1 and other loci, has established and FREELING1987; HARDEMANand CHANDLER that transposable elements, called Mu transposons, are 1989). The other four classes (Mu4, Mu5, Mu6 and likely to cause the high rate of mutation (STROMMER MuA) have been identified by their DNA homology et al. 1982; TAYLOR,CHANDLER and WALBOT1986; to the common -220-bp inverted terminal repeats O’REILLYet al. 1985). (TALBERT,PATTERSON and CHANDLER1989; QINand Nine different classes of Mu transposons have been ELLINGBOE1990; V. L. CHANDLER,unpublished). At distinguished on the basis of the sequence heteroge- least a few sequences characteristic of each class of neity within the element (BARKERet al. 1984; TAYLOR element are found in most non-Mutator standard and WALBOT1987; OISHIand FREELING1987; TAL- inbred lines of maize(CHANDLER, RIVIN and WALBOT BERT and CHANDLER1988; TALBERT,PATTERSON and 1986; TALBERT,PATTERSON and CHANDLER1989; V. CHANDLER1989; SCHNABLE,PETERSON and SAEDLER L. CHANDLERunpublished data). To date there is no 1989; VARAGONA,FLEENOR and WESSLER1987; QIN evidence that any of these characterized Mu elements and ELLINGBOE1990; V. CHANDLER,unpublished). encode a transposase (LILLIS,SPIELMANN and SIMPSON All classes ofMu elements have in common the -220- 1985; ALLEMANand FREELING1986; see TALBERT, base pair (bp) terminal inverted repeats, but different PATTERSONand CHANDLER1989 for discussion), al- classeshave distinct unrelated internal sequences. though, recently a MuA element has been shown to Memberswithin the same class share astrong se- be homologous to atranscript associated withMutator quence homology throughout the internal regions of activity (QINand ELLINCBOE1990). the elements. Furthermore, the described classes can Unlike the well-defined controlling element systems exist in multiple, heterogeneous copies and all ele- Ac-Ds and Spm, the genetic basis of Mutator activity ments shown to be capable of transposition generate appears to segregate in a non-Mendelian or multigenic manner (ROBERTSON 1978; WALBOT1986). When ’ Present address: Dekalb Plant Genetics, Eastern Point Rd.,Groton, active Mutator stocks are outcrossed to standard non- Connecticut 06340. Mutator lines, the typical result is that 90% of the

Genetics 149: 261-270 (September, 1991) 262 P. Chomet et al. progeny retain and 10% lose Mutator activity. How- Progeny segregating for the MuRl element(s) was obtained ever, these percentages are variable and simple men- from the cross: delian ratios reflecting the segregation of only one or +MuRl; al-mum2 Sh2/al sh2 X a1 sh2/al sh2; -MuRl two regulatory genes are rarely observed. Mutator- or inactive progeny are simultaneously altered relative to Mutator-active progeny with.respect to a number +MuRl; al-mum2 Sh2/al sh2 of phenotypesassociated with the Mutator system. X al-mum2 ShPlal sh2; -MuRl. First, the forward mutation rate is reduced, the ele- A Mutator stock, with a typical number of Mu1 elements ments are somatically stable and thehigh copynumber (20-60) was used to obtain new shl mutants. Details of the of Mu elements is not maintained upon outcrossing stock construction were previously described (HARDEMAN (ROBERTSONet al. 1985; BENNETZEN1987; WALBOT and CHANDLER1989). Eight shl mutants isolated at a fre- and WARREN1988), indicating Mu elements are no quency of 1.1 X were obtained (K. J. HARDEMANand longer transposing. Second, certain nucleotides in V. L. CHANDLER,unpublished). The shl mutant relevant to this work was designated shl-A83 and contained a 4.0-kb the ends of Mu transposonsbecome methylated insertion in the shl gene. (CHANDLERand WALBOT1986; BENNETZEN1987; Nucleic acid samples:Maize DNA from leaves was puri- BENNETZEN,BROWN and SPRINGER1988). Third, ex- fied according to CONE(1989). RNA was isolated from trachromosomal Mu circles associated with Mutator leaves and immature ears by the guanidinium thiocyanate activity are not detectable (SUNDARESANand FREEL- method and the sodium dodecyl sulfate (SDS) method (Au- SUBEL et al. 1987). Plasmids containing Mul, Mul.7, Mu4 ING 1987). Finally, a reporter allele that is normally and Mu5 probes were previously described (TALBERT,PAT- suppressed by Mutator activity and expresses a mutant TERSON and CHANDLER1989). A plasmid containing the phenotype, expresses the wild-type phenotype in Mu- Mu3 element was obtained from K. OISHI and a plasmid inactive lines(MARTIENSSEN et al. 1990). In some, but containing Mu8 was obtained from S. WESSLER.Internal fragments of the Mu6 and Mu7 elements were from D. not all cases, the loss of Mutator activity can be re- TURKS.The actin probe was obtained from R. MEACHER versed by crossing to Mutator-active plants (ROBERT- (MEACHERet al. 1983). A KfnI fragment from the 5' end SON et al. 1985; BENNETZEN1987; WALBOT1986). of the A1 gene was obtained from J. BROWN(O'REILLY et ROBERTSONand STINARD(1989) reported an un- al. 1985). usual Mutator stock where the control of Mu1 excision Cloning: An -8.5-kb sequence from the shl-A83 allele was cloned by ligating size-fractionated BclI-digested maize 1: segregated 1 in a testcross, suggesting the presence DNA into theBamHI site of the XBv2 vector (obtained from of a single regulator locus. Using a similar line con- N. MURRAY,University of Edinburgh). After in vitro pack- structed independently inthis laboratory we have aging (HOHN 1979), the phage were plated and screened identified a Mu transposon with regulatory activity (BENTONand DAVIS1977) using a Shl clone, p17.6, ob- and designated it Mu-Regulator1 (MuRI). Wehave tained from C. HANNAH.An 8-kb PstI fragment containing the entire Mu insertion, shl sequences, and -1.8 kb of the demonstrated that this gene is itself a transposon and X vector was subcloned into a pTZl8 vector (United States is necessary to confer activity on different classes of Biochemical). The site of insertion of the element in the shl Mu elements (this report; P. CHOMET,D. LISCHand allele was determined by sequencing the two SstI fragments M. FREELING,unpublished data). (subcloned into a pTZ 18vector) which contained homology We have identified and cloned two members of a to both the Mu-termini and Shl. Both strands were se- quenced by the dideoxy chain termination method (SANGER, new class ofMu elements that have not been described NICKLENand COULSEN1977) using the 17-mer universal before. These elements have the Mu terminal inverted primer (Pharmacia). An internal EcoRI-BamHI fragment of repeats, and hybridizewith the Mu5 element's ex- the element at shl-A83 was designated Mu*. tended terminal repeats. However, their internal se- A BclI clone containing most of the MuRl element was quences do not hybridize with those of Mul, Mul.7, obtained by size-fractionating BclI digested DNA from a plant carrying MuRl. This DNA was ligated into theBamHI Mu3, Mu4, Mu5, Mu6, Mu7or Mu8. In this paper we site of XEMBLS (purchased from Stratagene Inc.). The demonstrate that a member of this new class of Mu phage were in vitro packaged, plated and screened using elements is the MuRl transposon, a regulator for the the Mu* probe. A clone carrying the MuRl-diagnostic 1.4- multiclass Mutator system. In addition, we show that kb internal XbaI fragment was restriction mapped further. the MuR elements are homologous to two transcripts Southern and Northern blot hybridizations:Maize DNA was digested for 4 hr with a fourfold excess of restriction present in plants carrying MuRI. enzymes (conditions according to suppliers), electrophoretically separated through agarose gels, blotted to Duralon MATERIALS AND METHODS UV (Stratagene Inc.) and was UV cross-linked to the membrane using a Stratalinker apparatus (Stratagene Inc.) ac- Maize lines: The al-mum2 allele was originally isolated cording to the manufacturer. Electrophoresis of RNA was by D. S. ROBERTSONand the line containing al-mum2 in this through formaldehyde gels (AUSUBELet al. 1987). Probes study was originally obtained from S. DELLAPORTA.The al- were prepared by the random priming method using a dt sh2 tester line was obtained from B. MCCLINTOCK.The Prime-It kit (Stratagene Inc.) and "P-radiolabeled dCTP al-dt allele has a stable, colorless phenotype in the absence (Du Pont). Filters were prehybridized at 65" in 6 X SSC, 10 of Dt. All lines used inthis report lacked the Dt element and mM Tris-HCI, pH 7.5, 1 mM EDTA, 5 X Denhardt's solu- for this reason we simply refer to the a-dt allele as al. tion, 20 mM sodium phosphate, pH 7.2, 0.2 mg/ml dena- Control of the Mutator System 263

FIGURE 1.-Tasselbranch (A) and seed phenotypes (B) of the al-mum2 allele in the presence and absence of MuRl. Plants carryingMuRl show small red clonal sectors of tissue on a green background. Plants lacking MuRl express red anthocyanin color (the intensity of red is dependent on genetic background). The suppression of the al- mum2 allele is not readily apparentin the aleurone layer of the kernels. The genotypes of the plants and seeds shown are listed between each structure, respectively.

tured salmon sperm DNA, and 10% sarcosine and hybrid- some containing the reporter allele it was linked to ized in the same solution with the denatured probe for 6- Shrunken2 (Sh2)which produces plump kernels. The 12 hr. The filters were washed in 0.2 X SSPE, 0.1% SDS at recessive allele, sh2, was linked to astable recessivea1 65" for 1.5 hrwith three wash changes and exposed to Kodak X-Omat film with a Du Pont Cronex Lightning-Plus allele in the non-Mutator tester stock (see MATERIALS intensifying screen at -80" for varying lengths of time. AND METHODS for a furtherdescription of this allele). AI and Sh2 are closely linked with only 0.2%recom- RESULTS bination. Testcrosses (as shown in MATERIALS AND METHODS) Mutator activity is controlled by the MuRl trans- generated ears segregating 50% spotted, plump ker- posable element:To detect Mutator regulatory activ- nels and50% nonspotted, plump kernels, thereby ity we used al-mum2, an allele of the anthocyanin- defining a locus (MuRI)which regulated somatic Mu- conditioning gene AI. This mutant allele contains a tator activity. Most ears from crosses of thistype MuZ insertion in the 5' promoter region of the AI segregated for MuRI over a five generation period. gene (O'REILLYet al. 1985). In theabsence of Mutator Table 1 gives data for two consecutive generations of activity this allele produces a pale colored (nonspotted) such crosses. These results suggested the presence of seed, whereas in the presence of Mutator activity, the a single locus which was unlinked to al-mum2, and allele produces a spotted phenotype (Figure 1B). To was responsible for somatic excision of Mul. independently follow the segregation of the chromo- MuRl activity was also necessaryfor thesuppression 264 P. Chomet et al.

TABLE I & Segregation data for families with one element MuRl sh-83 dMuR .... S: E: i 5 X: .r H kc...i..S No. Sh2 kernels MuRl genomic - MuRl 5: E x 51 7 8% I1 I Plant No. parent" Mutable Pale X2 h Bd MuRl 994- 1 F 85 75 0.625 FIGURE2.-Restriction map comparison of the shl-AR? dMuR 994-1 M 68 34 1 1.333' element, a genomic MuRI element, and the Bcll clone containing 994-2 F 60 72 1.091 most of MuRI. The location of the Mu-termini are indicated as a 994-3 F 64 60 0.129 gray section and the approximatelocation of the Mu5 homologous 994-4 F 68 53 1.860 sequences are indicated as the cleared section of the dMuR in shl- 994-5 F 55 42 1.742 AR?. The dMuR is the only element confirmedto contain the Mu5 994-6 F 50 51 0.010 repeats; we havenot tested the Bcll clone orMuRl genomic 994-6 M 47 50 0.093 sequences for hybridization to the Mu5 repeat. The Mu* fragment 994-7 F 89 55 8.028' is diagrammed above the map. Restriction fragment sizes of each 994-7 M 39 45 0.429 elementwere directly compared on the same agarose gel. The 994-8 F 24 17 1.195 genomic restriction map of MuRl was generated from southern 994-9 F 51 67 2. I69 blots of complete digests of genomic DNA. Dashed lines between 994-9 M 55 61 0.310 the maps indicate the shared restriction sites and their positions. 994-1 0 F 65 70 0.185 The nine bp target site duplication of Shl sequence caused by the 994-10 M 59 54 0.221 dMuR element is: AGAGAAACC, which includes the Sh1 transcrip- 1277-2 F 94 61 7.026' tion start site (WERRet al. 1985). The restriction enzyme sites are 1277-2 M 61 50 1.090 abbreviated as follows: B, BamHl; Bc, Bcll; E, EcoRl; H, Hindlll; 1277-1 1 F 74 74 0.000 S, Sstl; X, Xbal. 1277-14 F 87 65 3.184 1277-14 M 107 85 3.000 with the complete internal probes forthe Mul, Mul.7, 1277-17 F 113 110 0.040 1277-20 M 98 103 0.124 Mu3, Mu4 and Mu8 classes of elements. It also failed to hybridize to small internal probesof Mu6 and Mu7. MuRl carrying parent was used as either a male (M) or female (F) in the cross as outlined in MATERIALS AND METHODS. However, asour Mu6 and Mu7 specific probes do not ' Fars with a significantly different x' value at the 0.05 level for represent the entire internal sequences of these ele- the segregation of one MuRl element. ments we can not completely eliminate the possibility of limited sequence homology between the new ele- of A1 expression mediated by the Mu1 element inser- ment andMu6 or Mu7. tion in al-mum2. In plants lacking MuRl activity (as &segregation of the MuRl locuswith a DNA assayed by nonspotted seed) we observed a strong red fragment homologous to the new Mu transposon: pigmentation in a variety of tissues whereas plants To determine whether thenew Mu transposon might carrying MuRl expressed a mutable pattern of red be related to theMuRl locus, siblings segregating for somatic sectors over a green background (Figure1 A). a single MuRl element were examined. An internal A new Mu transposonsequence identifies the EcoRI-BamHI fragment of the new element, desig- MuRl locus: Most transposon systems are character- nated as theMu* probe (Figure 2), was hybridized to ized by a regulatory element that encodes transpos-a Southern blotscontaining HindIII-digested DNA ase.In general, the regulatory elements are larger from 19siblings, 10 with MuRl activity and 9 without than their defective derivatives. At the time this study MuRl activity (see Figure 1 for reporter gene phe- was initiated, the characterized Mu elements ranged notypes). As shown in Figure 3, the Mu* probe de- in size from 1 to 2 kilobases (kb). Thus, we were tected multiple restriction fragments in all samples. interested in characterizing several larger Mu inser- This result was expected since the otherclasses of Mu tions (4-5 kb) at the Shrunkenl (Shl)locus that were transposons are present in multiple, dispersed copies potential candidates for regulatory elementsbased on in all Mutator-active maize genomes. Importantly, a their size. Adescription of the stocks and mutant strongly hybridizing 1 I-kb fragment was found in the isolation is outlined in MATERIALS AND METHODS. ten plants with MuRl activity and was absent in the Southern blot analyses were used to map the inser- plants lacking MuRl activity. The same cosegregation tions relative to Shl restrictions sites. We cloned one result was obtained in the analysis of other related insertion (shl-A83)and a restriction map of this ele- families. We have examined a total of 94 progeny ment is shown in Figure 2. This insertion hybridized segregatingfor MuRl and observed zero recombi- to the Mu1 terminal repeat probe indicating that this nants of the MuRl element andthis Hind111 fragment new insertion is amember of the Mutator family. (or acomparable 1.4-kb cosegregating XbaI frag- Furthermore, two fragments near the ends of the ment). Thus, the Mu* probe recognized a restriction insertion had homology to the 138-bp extended in- fragmentthat is tightly linked to the MuRl locus. vertedrepeats of Mu5 (TALBERT, PATTERSONand Although a few of the other Mu* cross-hybridizing CHANDLER1989). The insertion failed to hybridize fragments segregated in these populations, no other Control of the Mutator System 265 TABLE 2 MuRl segregation showingcopy number changes

No. Sh2 kernels MuRI No. Plant No. parent" Mutable Pale MuRl* 994- 1 F75 85 1 994-1' 34M 68 2 d 11 kb + 1 176A-1 F 5 98 - 1 176A-2 36F 93 2 1 176A-2' 1 M 89 13 3 1 176A-3 1 F 20 97 3 1 176A-4 1 F 20 69 2 1176B-1 63F 64 1 1176B-2 56F 56 1 1176E3 62F 70 1 1 176B-4 F 30 40 1 4 8A-1 F 37 23 1 4 8A-1 M 35 65 -d 8A-2 F 3 71 8 4 8A-3 F 2 94 24 8A-3 M 73 29 2 8A-6 F 43 20 2 FIGURE3.--Southern blot analysis of DNA from plants segregat- 8A-6 M 266 19 ing for MuRI activity. DNA from plants containing (+MuRI)and 8A-7 M 75 30 2 lacking (-MuRI) was digested with HindIII, electrophoretically 8B- 1 F 55 42 1 separated on a0.8% agarose gel, blotted and hybridized to the Mu* 8B- 1 M 57 55 1 d probe. An 1 1-kb fragment (marked by an arrow) is present only in 8B-4 M 40 76 - samples from MuRI-containing plants. 8B-6 F 38 41 1 8B-6 M 48 43 1 homologous fragment cosegregated with MuRI. 8B-7 F 49 44 1 Duplicative transpositionof MuRZ was associated 8 B-7 M43 56 1 with duplication of a Mu* hybridizing fragment: a MuRI carrying parent was used as either a male (M) or female Typically,most crosses of plants segregating for a (F) in the cross as outlined in MATERIALS AND METHODS. ' Minimum number of independently segregating MuRl to ob- single MuRl locus gave rise to mutable kernels which tain a nonsignificant x* value at the 0.05 level. showed similar frequencies of spotting. However, in ' Plant 994-1 M was the MuRl parent for family 1 176. Family some testcrosses, including those in which plant 994- 1 176 individuals marked A arose from heavily spotted kernels, and individuals marked with a B arose from medium spotted kernels. 1 was the Mutator male parent, an excess of spotted, The observed ratio is significantly different at the 0.05 signifi- plump kernels was observed (Tables 1 and 2) and cance level from the expected ratios for all possible numbers of approximately half of the spotted kernels were more independently segregating MuRI. ' Plant 1176A-2M was the MuRl parent for family 8. Family 8 frequently sectored. Previous observations indicated individuals marked A arose from heavily spotted kernels and indi- that increased copies ofMuRl in the endosperm pro- viduals marked with a B arose from medium spotted kernels. duced anincrease in the densityof the al-mum2 mediated spotting pattern in these lines (P. CHOMET, kernels displaying mediumspot densities allcontinued D. LISCH and M. FREELING,unpublished). We rea- to segregate 1 spotted:l nonspotted plump kernels soned that the increased spotting in the progeny of (Table 2, 1 176B family). From the data presented in plant 994-1 was due to an increase in MuRI copy Table 1, we can also infer that a germinal duplicative number. Fertilization by the gametes containing two transposition event to an unlinked site occurred in MuRI copies wouldbe initially detected by a frequent approximately 20% of the plants whichoriginally spotting pattern in the endosperm. The unusually carried one MuRI element. high proportion of spotted, plump kernels might have This genetic material provided the means to test resulted from a premeiotic duplicative transposition unequivocally whether the MuRI transposon hybrid- of MuRI during tassel development of plant 994-1, as ized to the Mu* probe. DNA samples from sibling this was observed in only the male transmitted ga- plants deduced to carry either one or two MuRI loci metes of the plant (Table l). by genetic analysis, thegrandparent plant (994-I), To testthis hypothesis, plants weregrown from and thetester line were digested with HindIII, South- these heavily spotted kernels and testcrossed with a1 ern blotted and hybridized with the Mu* probe. The sh2 plants. Ears resulting from these crosses (1 176A expected 11-kb fragment associated with MuRI was family) often segregated 3 spotted:l nonspotted, present (Figure 4; dark arrow) in the grandparent plump kernels consistent withthe presence of a second plant (994-1) but was absent in the plantslacking nonlinked regulator gene. Sibling plants grown from activity and was absent in the a1 sh2 homozygote 266 P. Chomet et al.

+ MuRl - MuRl 1022111- I It I

4-4.5 kb 4.5 kb 1

FIGURE4.-Identification of a unique Mu* hybridizing fragment FIGURE.5.-.l'he novel 4..5-kI> Mu* hybr-idizing fragment co- after duplicative transposition of MuRI. DNA from plants with segregates with MuRl activity. DNA from MuRl-segreg;lting prog- either two MuRI copies (plants 8A-3 and 8A-6), one MuRl copy eny of a plant carrying the new 4.5-kb fragment (8B-6) was digested (plants 8B-1,8B-6 and 8B-7). or zero MuRl copies (a1 sh2 tester with Hindlll, electrophoretically separated on a 1.0% agarose gel, line) was digested with HindIII. electrophoretically separated on a blotted and hybridized to the Mu* probe. The arrow points at the 0.8% agarose gel, blotted and hybridized to the Mu* probe. See segregating 4.5-kb fragment. Table 2 forgenetic analyses of these plants. The open arrow points at the new 4.5-kb fragment found in both DNA samples from plants cloned from shl-A83 as a derivative of MuR (dMuR). carrying two MuRl copies and in plants 8B-1 and 8B-6 which has We cloned a BcZI fragment containing most of a one MuRl element. The closed arrow indicates the original 1 1-kb band. MuRl element from a line which segregatedfor a single element. Restriction analysis comparing MuRI tester (known to genetically lack Mutator activity). In in the BclI clone, the dMuR element at Shl and a all the testcross progeny deduced to carry two MuRI proposed genomic restriction map of MuRI is shown genes by genetic analysis, a unique 4.5-kb fragment in Figure 2. Genomic mapping data indicated that in was detected (Figure 4; hollow arrow) in addition to our segregating lines the MuRI element had unique the original 11 kbfragment and the other Mu*- internal restriction fragments that were not shared by homologous fragments. This new 4.5-kb fragment was other MuR homologous sequences, such as a 1.4-kb not present in the 994 family (data not shown). In XbaI and a 1.9-kb EcoRI-Hind111 fragment (see Figure progeny plants segregating for only oneelement, 2). As shown by the restriction map, the MuRl clone either the original 1 1-kb or the new 4.5-kb band was contains all such restriction sites so farexamined. present, but not both (Figure4). Interestingly,restriction analysis of the dMuR ele- If the new 4.5-kb fragment was derived froma ment, cloned from an unrelated Mutator line, dem- duplication of the original MuRI locus, it too should onstrated that it is very similar to MuR1 except for a confer MuRl genetic activity. Progeny analysis of 40 putative deletion of approximately 0.9 kb near one individuals from a plant carryingonly the 4.5-kb Mu*- end of the element. Thisregion encompasses the BcZI homologous fragment (plant 8B-6) showed that this and oneof the HindIIIsites found in MuRI. The 1.4- fragment cosegregated with MuRI activity as was pre- kb XbaI and a 1.9-kb EcoRI-Hind111 internalfrag- viously shown for the 1 1-kb fragment.A subset ofthis ments are unique only in our MuRl segregating lines segregating population is shown in Figure 5. There- and are notlikely unique in other unrelated Mutator fore, either the original (11-kb HindIII) or the newly lines since the sh-83dMuR element contains these re- transposed (4.5-kb HindIII) MuRI gene was sufficient striction fragments. to confer somatic excision activity to Mul. Taken Somatic loss of Mutator activity reflects the loss togetherthese results demonstratethat the Mu* of the 1 1-kb MuRl fragment: Somatic sectors of pale probe is homologous to a regulatory element, MuRI, red tissue on a background of green tissue appeared which is itself a transposable element. Furthermore, in plants heterozygous for al-mum2 and carrying a one MuRI transposon, independent of its chromo- single MuRI locus. Those sectors represented losses somal location, is necessary to encode Mutator activ- of MuRl activity by two criteria:(1) the pale red ity. We will referto the transposableelement we phenotype within the sectors is the same as that of Control of the Mutator System 267 from a plant that contained the 11-kb HindIII frag- B ment. This red sector, which lacked MuRI activity, also lacked the 11-kb fragment. Similar results were obtained from plants which carried the 4.5-kb MuRl fragment. The adjoining leaf tissue which continued to show MuR1 activity contained theappropriate HindIII fragment, serving as a control. All the other - MuR-homologous fragments appeared in all samples independent of the activity of MuRI with the excep !.lkb") Yyr tion of a new MuR hybridizing fragment which appeared in the DNA from the red sector. The appearance of a new fragment has been observed in some, but not all, of the other red sectors examined (data not shown). It is unclear if this fragment represents . .. an alteration of the original MuRl fragment or is an alteration or duplication of another MuR sequence. Similar results, showing loss of the MuRI-containing fragment, were obtained using several other restriction enzymes suggesting the 1 l-kb fragment was lost and not absent due to incomplete digestion of the 7 kb DNA. Furthermore, Southern analysis of DNA from six independentsectors indicated that the unique, Hind I II internal EcoRI-Hind111 fragment of MuRl was lacking Mu' probe in all six red sector samples (data not shown). The consistent lack of the MuRl fragment in mutator-loss FIGURE6.-Loss of MuRI activity in a somatic sector correlates with the loss of the 1 I-kb MuRI fragment.DNA was prepared from sectorsprovides independent conformation that a red sector of sheath tissue (red sector) and an adjoining region of MuRl is necessary for mutator activity. tissue whichstill showed MuRl activity (mutable). With the red Transcripts homologous to Mu* areassociated sector, effortswere made to remove the nonpigmented tissue before with MuRl: To determine whether there were tran- isolating DNA. Panel A shows a Southern blot hybridized to theAI probe of the DNA samples digested with HinfI, a methylation scripts homologous to MuR that correlated with both sensitive enzyme. The 2.1-kb fragment represents the entire Mu1 MuRI and Mutator activity in general, we performed element methylated at the Hinflsites and flanked by AI sequence. Northern blot hybridizations. Figure 7 shows the re- The 0.7-kb fragment represents complete digestion at the Hinfl sults of hybridizationof the Mu* probe to a blot site in the Mu1 terminal inverted repeat. The otherAI hybridizing carrying poly(A)+ RNA from four siblings, two con- fragments are from regions outside of the Mu1 insertion in a1- mum2 and from the other allele and are found in both samples. taining (+MuRI) and two lacking MuRI(-MuRI) Panel B is a Southern analysis of a different butcomparable sector. from a population segregating 1 :1 for MuRI. Two The DNA was digested with HindIII, electrophoretically separated transcripts, migrating at approximately 2.5 and 0.9 on a 0.8% gel, and hybridized to the Mu* probe. The 1 I-kb MuRI kb were homologous to the Mu* probe, and were fragment (marked by the arrow) is present in the sample from the mutable tissue but is absent in the DNA from the red sector. An present in segregants carryingMuRI only. These tran- additional, larger MuR-homologous fragment is detected in the red scripts may represent a MuRI encoded transcript in- sector. volved in MuRl activity, or represent transcripts from defective MuR elements thatare regulated by Mutator plants carrying the al-mum2 allele but lacking MuRI activity. We have examined otherMutator-active, Mu- (Figure 1) and (2) HinfI digestion of the DNA from tator-inactive and non-Mutator stocks from diverse tissue of many independent sectors revealed theMu1 backgrounds and found the2.5-kb transcript and the element inserted at al-mum2 was methylated in its smaller 0.9-kb transcript in all Mutator-active stocks inverted terminal repeats (Figure 6A). CHANDLERand examined to date (data not shown), supporting the WALBOT(1 986) and BENNETZEN(1 987)have shown hypothesis thatboth poly(A)+ RNA species are in- that modification of methylation sensitive sites in Mu1 volved in Mutator activity. ends correlates with the lack of Mutator activity. To further analyze the putative transcriptionalunits Given that MuRl activity was lacking in these sectors, of MuRI, two additional probes flanking MU*, Mu we investigated what happened to the MuRl fragment E-S and Mu B-H (see Figure 7), were also hybridized in these sectors. to similar Northern blots. The resultingdata are DNA from pale red sectors and adjoining mutable summarized in Figure 7. Whereas Mu* hybridized to tissue was analyzed by Southern blotting. One exam- both the 2.5-kb and the 0.9-kb transcripts, the Mu B- ple is shown in Figure 6B. The sector was obtained H probe hybridized only to the larger transcript. No P. Chomet et al. indicates the dMuR element is missing 0.9-kb relative to MuRl (Figure 2). Many of the dMuR elements are likely to be rearrangement, insertion or deletion var- iants of each other much like other nonautonomous elements of controlling element systems such as Dis- sociation (Ds)or dSpm (FEDOROFF,WESSLER and SHURE 1983; PEREIRAet al. 1985). At least three different families of Ds elements have been described, each 0.9 kb + family containing multiple members that are heterogeneous in size (SACHS et al. 1983; BURRand BURR 1981 ; FEDOROFF, WESSLERand SHURE 1983; DORINC et al. 1984). Molecular analyses of other dMuR elements we have cloned may show their familial rela- tionships. Mutator-active stocks contain MuR-homologous transcripts that are absentin non-Mutator and Muta- tor-inactive stocks. Presence of these transcripts correlates with MuRI activity. These transcripts do not hybridize with the Mu termini, andare therefore unlikely to result from the simple readthrough from a gene into an endogenousdMuR element. Thus, our working hypothesis is that one or both of these tran- Transcripts (kb): 4 4 4 2.5 ~ + t scripts encodes aproduct involved in Mutator activity. Isolation and characterization of thesetranscripts 0.9 - ' . should reveal clues as to their role in Mutator activity. FIGURE7.-ldentification of Mu*-homologous transcripts asso- Unlike other two-element systems in maize, each of ciated with Mutator activity. A sample of 1 pg of poly(A)+ RNA whichusually segregate for a single regulatory ele- from immature ears of containing lanes) and MuRI (+MuRI MuRI ment, ROBERTSON'SMutator system does so only very lacking (-MuRI lanes) plants was separated on a 1 % formaldehyde agarose gel and hybridized to the Mu* probe. The 2.5-kb and0.9- rarely. Upon outcrossing, most progeny from a Mu- kb RNA species are indicated by arrows. The Mu* probe was tator plant receive Mutator activity. To explain the washed off the membrane (according to the manufacturer) and the unusual segregation, ROBERTSON(1978) and BEN- blot was hybridized with the maize Actin probe (lower panel) to NETZEN (1984) suggested that multiple copies of a show each lane contained approximately equal amounts of intact RNA. Distinct restriction fragments of dMuR were used as probes regulatory element would have to exist and such cop to additional Northern blots and the results are summarized in the ies would have to undergo premeiotic amplification diagram below the blot. A "+" indicates that the particular restric- to be passed on to most progeny.Although Mu1 tion fragment probe detected the indicated transcript. A "-" indi- transposons behave according to these expectations cates no signal was detected. (ALLEMANand FREELINC1986), Mu1 does not regu- hybridization signal on Northern blots was detected late the Mu system (LILLIS,SPIELMANN andSIMPSON with the Mu E-S probe. Therefore, it is likely that the 1985; P. Chomet, unpublished). Our data show that 0.9-kb transcriptionalunit is contained within the MuRl does undergo premeiotic duplicative transpo- limits of the Mu* sequence and that the 2.5-kb tran- sition events which result in the increase in MuRI scriptional unit extends to the right side of that se- copy number. Although the frequency of duplication quence. we observed may not completely explain the unusual segregation patterns of Mutator activity, it is possible that multiple MuRI elementsact synergistically to DISCUSSION increase their rate of duplication. This possibility is A member of a previously unidentified class of Mu currently being tested. elements was cloned from a Mu-induced shrunken1 WALBOT(1 986) has argued that the segregation of allele. We have shown that this 4.0 kb element molec- Mutator activity is such that simple, mendelian regu- ularly identifies the genetically characterized MuRI latory genes cannot bethe proximatecause of the loss element, which regulates the activity of Mu elements. of Mutator activity in her lines. This conclusion There are multiple fragments homologous to MuRI stemmed from observations in which the active state in all maize plants examined,even plants lacking of Mutator, as assayed by somatic instability of Mul, MuRI or Mutator activity. The 4.0-kb element cloned can become weakly active with only a few kernels from shl-A83 appears to be incomplete, as a restriction showing somatic excisions, or the active state can analysis comparison between this dMuR and MuRl become inactive through repeated self pollinations. It Control of the Mutator System 269 is likely that these data are not explained simply by analyses will demonstrate if the sequence of MuRl is meiotic segregation of regulatory elements. We do homologous to that of Cy. know that MuRl canfunctionally replace Mutator Recently, a ninth class ofMutator elements has been activity when Mutator activity was lost via outcrossing described. One member of the family, designated (P. CHOMET,D. LISCHand M. FREELING,unpublished MuA, was cloned and shown to be homologous to a data). Furthermore, weshow here that inactivity in 3.5-kb transcript associated withMutator activity (QIN our lines is primarily due to the lack of the MuRl and ELLINGBOE1990). Although no evidenceindi- element, although we do note exceptional plants cates that MuA is the autonomous element, it is pos- whichgive rise to non-mendelian lossof Mutator sible that MuA and MuRl are members of the same activity (see Table 2). Therefore, it is possible that classof Mu elements, as they have similar but not genetic loci, such as MuRl, are responsible for the identical restriction maps. The reason for the discrep- segregation of Mutator activity but the unusual be- ancy in size between the reported 3.5-kb transcript haviors may be caused by stable or metastable second- andthe 2.5-kb transcript we observe is not clear. ary modifications (such as DNAmethylation) to these Different electrophoresis conditions and RNA mark- regulating loci rendering them functionless. ers might have led to different size estimations. Alter- The simultaneous occurrence of the loss of Mutator natively, the two elements may be hybridizing to dif- activity, the methylation of Mul, and the loss of the ferent transcripts. A direct comparison between the MuRl restriction fragment in somatic sectors, suggests two elements will clarify the differences. that the loss of MuRl is associated with the gain of The recent insertion of the dMuR element into Shl Mu1 methylation. Furthermore, loss ofMuRl through caused a 9-bp host site direct duplication. All other mendelian segregation is alsoassociated with Mu1 recently transposed Mutator elements also produced a methylation (P. CHOMET,D. LISCHand M. FREELING, 9-bp duplication (BARKERet al. 1984; CHEN et al. unpublished data). One possibility for such results is 1987; SCHNABLE,PETERSON and SAEDLER1989). Such that Mu1 methylation is the consequence of the loss host site duplications are a common feature of trans- of MuRl in these instances. It would follow then that posable elements and the extent of the duplication is Mu1 methylation is the “ground state” in the absence specific for each family of interacting elements. Since of MuRl activity. By this reasoning, the presence of all elements within a family are under the control of MuRl product(s) should block Mu1 methylation. A a common transposase, the host-site duplication length DNA footprint in the Mul-end sequence associated is likely a reflection of the shared transposase-me- with Mutator activity hasbeen detected (Z. ZHAO and diated insertion mechanism. This suggests the trans- V. SUNDARESAN,personal communication). We sug- position of MuR elements, along with the other Mu- gest that MuRl product(s) bind to these protected nucleotides. tator elements, is controlled by a similar mechanism, The mechanism underlying MuRl loss in somatic likely regulated by MuRl. sectors is unclear. It is possible that the element was Although Mu transposons are excellent insertional altered or rearranged during the transposition proc- mutagens in maize, it has been difficultto clone some ess. The appearance of a new fragment in some Mu-induced mutations due to the complexity of the of these sectors would be consistent with such an as- system. The identification of a regulatory element will sertion. The new fragment may represent a re- greatly facilitate the use of the Mutator system as a arrangement derivative at the MuRl location or at a transposon tag in maize. Furthermore, a demonstra- new genomic position. In either case, it is likely these tion of Mu transposition in transgenic systems would new fragments appearing in the red sectors are not prove to be a valuable toolfor insertional mutagenesis complete MuRl elements since analysis of independ- and cloning in other plants as well. Sequences homol- ent sectors indicated that aunique, internal restriction ogous to MuRI have been detected in other monocots fragment, associated with MuRl, was lacking in the (D. LISCH,unpublished data). An efficient transposon- red sectors (data not shown). More detailed analyses tagging system for major crops other than maize to furtherelucidate these alterations are underway. would be particularly valuable.

Genetic studies on the Cy controlling element sys- We thank Doctor A. BARKAN for herinitial observations using tem demonstrated that Cy is a transposable element the Mu* probe. We thank S. BELCHER,A. CHYTRY,R. LACHMAN- capable of regulating Mu1 and Mu7 (SCHNABLEand SINGH, Y. ORTEGA,V. POPADICand N. WALKER for technical PETERSON1988; SCHNABLE,PETERSON and SAEDLER assistance and R. K. DAWE, B.KLOECKNER, B. LANE andJ. VOGEL 1989). Cy can segregate as a single mendelianelement for critical reading of the manuscript. This work was supported by and has been observed to increase its copy number the National Science Foundation (DCB-87115 15) to M.F. and the National Institute of Health (GM 35971) to V.L.C. P.C. was (SCHNABLEand PETERSON1988). Presumably, this supported by a National Institute of Health Postdoctoral Training means that Cy is a regulatory element of the Mutator Grant (GM12723) and K.J.H. was supported by the Predoctoral system,functionally analogous to MuRI. Molecular National Institute of Health Training Grant (GM07413). 270 P. Chomet et al.

LITERATURECITED O’REILLY,C., N. S. SHEPHERD,A. PEREIRA,Z. SCHWARZ-SOMMER, I. BERTRAM,D. S. ROBERTSON,P. A. PETERSONand H. SAE- ALLEMAN,M., and M. FREELING,1986 The Mu transposable DLER, 1985 Molecular cloning of the a2 locus of Zeamays elements of maize: evidence for transposition and copy number using the transposable elements En and Mu. EMBO J. 4 591- regulation during development. Genetics 112: 107-1 19. 597. AUSUBEL,F. M., R. BRENT,R. E. KINGSTON, D. D. MOORE,J. S. PEREIRA,A., Z. SCHWARZ-SOMMER,A. GIERL, I. BERTRAM,P. A. SEIDMAN,J. A. SMITHand K. STRUHL,1987 Analysis of RNA PETERSONand H. SAEDLER,1985 Genetic and molecular by Northern hybridization, pp. 4.9.1-4.9.7 in Current Protocols analysis of the Enhancer (En) transposable element system of in Molecular Biology, edited by F. M. AUSUBEL.Wiley Intersci- Zea mays. EMBO J. 4: 17-23. ence, New York. QIN,M., and A. H. ELLINGBOE,1990 A transcript identified by BARKER,R. F., D. V. THOMPSON,D. R. TALBOT,J. SWANSONand MuA of maize is associated with Mutator activity. Mol. Gen. J. L. BENNETZEN,1984 Nucleotide sequence of the maize Genet. 224: 357-363. transposable element Mul. Nucleic Acids Res. 12: 5955-5967. ROBERTSON,D. S., 1978 Characterization of a mutator system in BENNETZEN,J. L., 1984 Transposable element Mu2is found in maize. Mutat. Res. 51: 21-28. multiple copies only in Robertson’s Mutator maize lines.J. Mol. ROBERTSON,D. S., and P. S. STINARD,1989 Genetic analysisof Appl. Genet. 2: 519-524. putative two-element systems regulating somatic mutability in BENNETZEN,J. L., 1987 Covalent DNA modification andthe Mutator-induced aleurone mutants of maize. Dev. Genet. 10 regulation ofMutator elementtransposition in maize. Mol.Gen. 482-506. Genet. 208: 45-5 1. ROBERTSON,D. S., P. S. STINARD,J. G. WHEELERand D.W. BENNETZEN,J. L., W.E. BROWNand P. S. SPRINGER,1988 The MORRIS, 1985 Genetic and molecular studies on germinal state of DNA modification within and flanking maize transpos- and somatic instability in Mutator-induced aleurone mutantsof able elements, pp. 237-250 in Plant TransposableElements, maize, pp. 3 17-33 1 in Plant Genetics, edited by M. FREELING. edited by 0. NELSON.Plenum Press, New York. Alan R. Liss, New York. BENTON,W. D., and R.W. DAVIS1977 Screening Xgt recombi- SACHS,M. M., W. J. PEACOCK,E. S. DENNISand W. L. GERLACH, nant clones by hybridization to single plaques in situ. Science 1983 MaizeAc-Ds controlling elements-a molecular view- 196 180-182. point. Maydica 28: 289-301. BURR,B., and F.A. BURR,1981 Controlling-element events at SANGER,F., S. NICKLENand A. R. COULSEN,1977 DNA sequenc- the Shrunken locus in maize. Genetics 98: 143-156. ing with chain terminating inhibitors. Proc. Natl. Acad. Sci. CHANDLER,V., C. RIVINand V. WALBOT,1986 Stable non-mu- USA 74 5463-5467. tator stocks of maize have sequences homologous to the Mu1 SCHNABLE,P. S., and P. A. PETERSON,1988 The Mutator-related transposable element. Genetics 114 1007-21. Cy transposable element of Zeamays L. behaves as a near- CHANDLER,V. L., and V. WALBOT,1986 DNA modifications of a Mendelian factor. Genetics 120 587-596. maize transposable element correlates with loss of activity. SCHNABLE,P. S., P. A. PETERSONand H. SAEDLER,1989 The Proc. Natl. Acad. Sci. USA 83: 1767-1771. bz-rcy allele of the Cy transposable element system of Zea mays CHEN,C. H., K. K. OISHI, B. KLOECKENER-GRUISSEMand M. contains a Mu-like element insertion. Mol. Gen. Genet. 217: FREELING, 1987 Organ-Specific Expression of maize Adhl is 459-463. altered after a Mu transposon insertion. Genetics 116 469- STROMMER,J. N., S. HAKE,J. L. BENNETZEN,W. C. TAYLORand 477. M. FREELING,1982 Regulatory mutants of the maize Adhl CONE,K., 1989 Yet another rapid plant DNA prep. Maize Genet. gene caused by DNA insertions. Nature 300 542-544. Coop. Newsl. 63: 68. SUNDARESAN,V., and M. FREELING,1987 An extrachromosomal DORING,H. P., M. FREELING,S. HAKE,M. A. JOHNS,R. KUNZE,A. form of the Mu transposon in maize. Proc. Natl. Acad. Sci. MERCKELBACH,F. SALAMINI and P. STARLINGER,1984 A Ds- USA 84 4924-4928. mutation of the Adhl gene in Zeamays L. Mol. Gen. Genet. TALBERT,L. E., and V. L. CHANDLER, 1988Characterization of 193: 199-204. a highly conserved sequence related to Mutator transposable FEDOROFF,N., S. WESSLERand M. SHURE, 1983 Isolation of the elements in maize. Mol. Biol. Evol. 5: 519-529. transposable maize controlling elements Ac and Ds. Cell 35 TALBERT,L. E., G. I. PATTERSONand V. CHANDLER,1989 Mu 235-242. transposable elements are structurally diverse and distributed HARDEMAN,K. J., and V. L. CHANDLER,1989 Characterization throughout the genus Zea. J. Mol. Evol. 29 28-39. of bzl mutants isolated from mutator stocks with high and low TAYLOR,L. P., V. L. CHANDLERand V. WALBOT,1986 Insertion numbers of Mu1 elements. Dev. Genet. 10460-472. of 1.4 kb Mu elements into the bronze1 gene ofZea mays L. HOHN,B., 1979 In vitro packaging of lambda and cosmid DNA. Maydica 31: 31-45. Methods Enzymol. 68: 299-309. TAYLOR,L. P., and V. WALBOT,1987 Isolation and characteriza- LILLIS,M., A. SPIELMANNand R. B. SIMPSON,1985 Transfer of a tion of a 1.7-kb transposable element from a mutator line of Robertson’s mutator element to tomato and tobacco cells, pp. maize. Genetics 117: 297-307. 213-224 in Plane Genetics, edited byM. FREELING.Alan R. VARAGONA,M., D. FLEENORand S. R. WESSLER,1987 Unique Ds Liss, New York. and Mu elements inserted in exons of the Waxy gene, pp. 38 1 MARTIENSSEN,R., A. BARKAN,W. C. TAYLORand M. FREELING, in Plant Transposable Elements, edited by 0. NELSON.Plenum 1990 Somatically heritable switches in the DNA modification Press, New York. WALBOT,V., 1986 Inheritance of Mutator activity in Zea mays as of Mu transposable elements monitored with a suppressible assayed by somatic instability of the bz2-mu1 allele. Genetics mutant in maize. Genes Dev. 4 331-343. 114: 1293-1312. MEAGHER,R. B., R. C. HIGHTOWER,D. M. SHAHand T. MOZER, WALBOT,V., and C. WARREN,1988 Regulation ofMu element 1983 Plant actin is encoded by diverse multi-gene families, copy number in maize lines with an active or inactive Mutator pp. 17 1-1 87 in Advances in Gene Technology: Molecular Genetics transposable element system. Mol. Gen. Genet. 211: 27-34. of Plants and Animals, edited by K. DOWNEY,R. VOELLMEY,F. WERR, W., W. B. FROMMER,C. MAAS and P. STARLINGER, AHMADand J. SHULTZ.Academic Press, New York. ‘1985 Structure of the sucrose synthase gene on chromosome OISHI,K., and M. FREELING,1987 A new Mu elementfrom a 9 ofZea mays L. EMBO J. 4: 1373-1380. Robertson’s Mutator line, pp. 289-291 in Plant Transposable Elements, edited by 0. NELSON.Plenum Press, New York. Communicating editor: B. BURR