J . Cell Sci. Suppl. 7, 123-138 (1987)Printed in Great Britain © The Company o f Biologists Limited 1987

123

ASPECTS OF THE A c /D s TRANSPOSABLE ELEMENT

SYSTEM IN MAIZE

W. J A M E S P E A C O C K , E L I Z A B E T H S. D E N N I S , E. J E A N

F I N N E G A N , T H O M A S A. P E T E R S O N a n d B R I A N H. T A Y L O R

CSIRO Division of Plant Industry, GPO Box 1600, Canberra, ACT 2601, Australia

S U M M A R Y

Studies of the Ac (Activator) transposable element provided the data which led Barbara

McClintock to postulate that certain segments of chromosomes could transpose to different

locations in the genome. McClintock also recognized the existence of Ds (Dissociation) elements

which could transpose, but only in the presence of a trans-acting Ac element elsewhere in the

genome. D N A sequences corresponding to Ds and Ac have now been identified, and an

understanding of many of the properties of these transposable elements in the maize genome has

been acquired in recent years.

It is known that cryptic Ac elements and members of at least two families of Ds elements occur in

the genome of all maize lines examined. Ds elements also occur in Teosinte and the more distantly

related Tripsacum. We discuss the possible origin of these elements and consider the mechanism of

activation of cryptic Ac elements. A recent molecular analysis of a transition of an Ac-derived D i ­

element back to an active Ac element suggests one molecular mechanism by which changes in the

activity state of Ac may occur.

Distinctive phenotypes created by controlling elements within a target gene have been shown to

be governed by the properties of the insertion element and the position of the insertion within the

gene. Genetic effects include modulation of gene expression, alteration of gene products, instability

of mutant phenotypes, deletion and duplication of chromosome segments and the production of

chromosome rearrangements. We describe an example where a Ds insertion generates an additional

intron in the A dhl gene which reduces gene expression through mRNA instability. We also discuss

an Ac-dependent modulation of P gene activity in glume and pericarp tissues of maize which may

be attributed to an alteration either in patterns of gene expression or the developmental biology of

the flower.

The molecular consequences of Ac and Ds insertions and excisions are known at the DN A

sequence level but little is known of the mechanism of transposition. An initial approach has been

to analyse Ac transcription. Preliminary results showing transcription of a limited region of Ac are

discussed. The corresponding upstream regions have been linked to the coding region of

chloramphenicol acetyltransferase (CAT) and show promoter activity following electroporation

into tobacco protoplasts.

I N T R O D U C T I O N

Maize transposable elements have at times been considered as biological

anomalies, and those studying them as participating in fringe science. Clearly such

opinions are no longer tenable. Geneticists have long sought to identify and

characterize those factors which control and modify the coordinated expression of the

thousands of genes which are needed to produce an organism, the functional unit of

biological organization. In the late 1940s Barbara McClintock identified certain

factors in maize which seemed to control gene expression; she termed these factors

controlling elements.

124 W. J . Peacock and others

McClintock (1951, 1956) recognized two fundamental genetic properties of the

maize controlling elements which she studied. First, the elements are capable of

profoundly altering the expression of specific genes with which they are associated.

Association of a controlling element with a gene can partially or completely inhibit

expression, resulting in a range of possible activities from null to near normal levels.

Expression of the affected gene can be altered in developmental time or tissue

specificity, producing new patterns of gene expression. The second fundamental

property of McClintock’s elements is their ability to transpose to new genomic sites.

At these new sites, the elements can again affect gene control.

Presently there are over a dozen different controlling element systems known in

maize. Each system may comprise several families, with each family containing

perhaps 50 or more individual elements. Freeling (1984) has suggested that up to

50 % or more of the maize genome is or was transposable. McClintock identified and

studied two such systems, namely the Spm system and the Ac/Ds system. In this

paper we shall consider primarily the Ac/Ds system; many of its characteristics apply

to other systems.

The Ac/Ds system comprises two types of elements which are distinguished by

their transpositional capabilities. Ac (Activator) elements transpose and also enable

Ds elements to transpose. Ds (Dissociation) elements transpose only when an Ac element is present; they cannot transpose without Ac, nor do they promote

transposition of other Ds elements. McClintock found that a Ds element could be

derived from Ac, suggesting that at least some Ds elements are defective versions of

Ac.Individual Ac and Ds elements have now been isolated and characterized, and

their descriptions have provided molecular explanations for some of the genetic

phenomena. It is now known that Ac and Ds elements are discrete genetic units with

characteristic structure and DNA sequence; reversion of Ac or Ds-induced mutants

is correlated with excision of the element. The effect of Ac and Ds elements on gene

expression results from an interaction of element and gene in a manner which is a

function of the sequence of each and their relative positions. It is still unclear

whether McClintock’s elements participate significantly in developmental control of

gene expression. However, studies of these elements and the genes with which they

are associated are providing an incisive view of the molecular events underlying

development.

S E Q U E N C E C H A R A C T E R I S T I C S OF Ac A N D D s E L E M E N T S

A number of examples of both Ac and Ds elements have been analysed at the

molecular level. Most of these elements were recovered from mutant loci originally

described by McClintock, although several are of more recent origin. Two Ac elements, as well as a number of Ds elements, have been sequenced in their entirety.

The Ac elements which have been sequenced were both originally detected as

instabilities in the waxy locUs of maize and were designated Ac wx-m7 and Ac wx- m9. Ac wx-m7 was unusual in that it appeared to cycle between an Ac form and a Ds

Ac/Ds tratisposable elements 125

Open reading frame map of Ac (wx—m9)232

132 k H = >

120

H in dW lI

i>104

BamUl

0 0-5

102

•Met

1-0

91

<=

1-5

294

2-0 2-5 3-0

105

3-5

$

4-0

168

4-5

167

485 , 128

< = 3

Fig. 1. Open reading frame map of Ac wx-m9. Open reading frames were determined by

computer analysis of sequence data obtained by Pohlman et al. (1984a,b). Only open

reading frames in excess of 75 codons are shown. The position of the first methionine

codon in each open reading frame is indicated by an asterisk. The 10 base pair terminal

inverted repeats are indicated by short arrows at the ends of the element.

form, while Ac wx-m9 continued to function as an Ac. Ac wx-m9 was cloned first

using a waxy gene probe, followed by Ac wx-m7, which was identified by

hybridization to both waxy DNA and a probe containing DNA from a Ds insertion

(Fedoroff et al. 1983; Behrens et al. 1984). The sequences of these elements were

found to be nearly identical (Pohlman et al. 1984a,b ; Muller-Neumann et al. 1984)

suggesting that the cycling phenotype of Ac wx-m7 was not sequence specific.

Indeed, a cycling derivative of Ac wx-m9 was discovered by Schwartz (Schwartz &

Dennis, 1986) and is described later in this review.

The Ac elements cloned from the waxy gene are 4563 bp in length and are

bounded by 10 bp indirect repeats, with an additional noncomplementary base on the

termini. The Ac insertions are flanked by 8 bp duplications of host DNA which were

apparently generated when the Ac elements entered the target DNA. Computer

analysis of the Ac sequence reveals the presence of a number of large open reading

frames, with those in excess of 75 codons shown in Fig. 1. At the present time the

significance of these open reading frames has not been established, although

sequence analysis of a Ds element which arose directly from /\c wx-m9 indicates that

a 194 bp deletion within the 485 codon ORF is sufficient to block Ac function

(Pohlman et al. 1984«). This result suggests that at least part of the 485 codon ORF

is translated into protein which is part of a trans-acting transposase or a modulator

thereof.

The internal sequence of Ac contains a number of direct and inverted repeats, the

functions of which are unknown. Electron micrographs of a segment of cloned Ac

wx-m7 DNA heteroduplexed to a cloned segment of the wx gene appeared to show a

stem structure of approximately 150 bp involving the ends of the Ac element

(Behrens et al. 1984). Subsequent computer analysis of the sequence of the Ac wx-

126 W. J . Peacock and others

m7 termini (Muller-Neumann et al. 1984) indicated that such a structure was

possible, although evidence indicating a role for this structure in vivo is lacking.

As mentioned previously, Ds elements can be generated by internal deletions of

Ac. Examples of this type of Ds include Ds wx-m6 (Pohlman et al. 1984a) and sh- m6233 (Week et al. 1984). sh-m6233 is a complex insertion consisting of one Ac deletion-derived Ds inserted in reverse orientation into a second, identical Ds. This

structure is also present in what appears to be an even more complex insertion in sh- m5933. In this locus, the double Ds insertion is repeated in deleted form at the

opposite end of a 30 kb insertion. Revertants of sh-m5933 have lost the 30 kb insert,

indicating that Ds elements can interact to move large segments of DNA (Courage-

Tebbe et al. 1983). Hybridization of the terminal regions of Ac to maize genomic

DN A indicates the presence of a large number of Ac-related sequences (Pohlman

et al. 1984a). Similar hybridizations with central regions of Ac revealed a much

lower multiplicity of sequences. This suggests that many of the hybridizing

sequences are Ac deletion-derived Ds elements, although other origins of these

sequences are possible.

A second class of Ds element is typified by the D sl element isolated from Adhl- Fm335 (Peacock et al. 1983; Sutton et al. 1984). D sl is inserted into the transcribed

leader sequence of the alcohol dehydrogenase 1 (A d h l) gene of maize, resulting in a

mutant phenotype with low ADH activity (Osterman & Schwartz, 1981). The insert

is 405 bp and does not hybridize to the Ac wx-m9 clone. Sequence analysis revealed

an 11 bp inverted repeat at the ends of the element, 10 bp of which are identical to

that of Ac, with an additional complementary nucleotide on the outermost ends. Like

Ac and the Ac-derived Ds elements, this element produces an 8 bp duplication of

host D N A upon insertion. It is genetically a Ds element in that there is a high

frequency of reversion to full ADH activity only when an Ac element is present in the

genome (Osterman & Schwartz, 1981). With the exception of the termini the

sequence of D sl is unrelated to Ac, suggesting that the 11 bp repeats are sufficient for

Ac-induced transposition.

Analysis of revertants of this D sl element provided the first data on the molecular

consequences of excision of this family of transposable elements (Peacock et al. 1983;

Sachs et al. 1983; Sutton et al. 1984). Schwartz obtained a number of reversions of

the mutant to normal levels of ADH activity which were cloned and sequenced across

the original site of insertion. In these revertants the 8 bp duplication is generally

retained, although in at least one instance a perfect excision has occurred (Dennis

et al. 1986). A common observation is that nucleotides directly adjacent to the site of

insertion have been deleted or changed to the complementary nucleotides (Sachs

et al. 1983; Pohlman et al. 1984a,b \ Week et al. 1984; Dennis et al. 1986). Models

which have been proposed to explain this observation suggest that because these

alterations do not always occur and are variable in nature, they are likely to arise

during DNA repair following excision (Peacock et al. 1984; Saedler & Nevers,

1985). Larger deletions are also associated with Ac/Ds excision (McClintock, 1950)

and may involve recombination with related sequences away from the site of

Ac/Ds transposable elements 111

insertion (Dennis et al. 1986) or another copy of the transposable element (Courage-

Tebbe et al. 1983).

The propensity of the Ds element sh-m5933 for generating chromosomal

rearrangements and breakages may be a consequence of its particular structure

(Dóring et al. 1984). Most Ac and Ds elements do not have a high frequency of

chromosomal breakage associated with their insertion/excision cycles. The occur­

rence of several elements of one particular family in different loci within the genome

could lead to chromosomal rearrangements by recombination between them. This

would be comparable to the translocation-generating events analysed by Mikus &

Petes (1982). In Drosophila it has been proposed that hot spots of chromosomal

rearrangement may be due to recombination between members of a dispersed family

of transposable elements (Engels & Preston, 1981).

F A M I L I E S OF D s E L E M E N T S I N M A I Z E A N D R E L A T E D S P E C I E S

There are a number of Ac and Ds related segments in the maize genome. Using

D sl as a probe, Sutton et al. (1984) reported 40—50 bands in Southern hybridiz­

ations. Fedoroff et al. (1983) reported a similar number of bands for the Ac-related

Ds family. These do not cross hybridize with the D sl segments. Segments which

hybridize at high stringency to the D sl element have also been found in Teosinte, the

immediate wild precursor of maize, and in Tripsacum dactyloides, a more distantly

related species (Gerlach et al. 1987).

Gerlach et al. (1986) cloned a number of these related segments and have shown

that in all cases the cloned fragments contained a D sl sequence of approximately

400 bp, with at least 90 °lo sequence homology to D s l . All but one of these segments

have the 11 bp inverted repeat termini characteristic of D sl and are bounded by

direct 8 bp repeats. The remaining element, D slO l, had a 10 bp inverted repeat at its

termini and was flanked by a duplication of 6bp rather than 8 bp. This suggests that

the length of the inverted repeat may influence the length of the staggered cut in the

genomic target. Other members of the D sl family have been cloned by Wessler et al. (1986) from the wx-m l mutation and by Schiefelbein & Nelson (personal communi­

cation) from the bz-wm mutation. There is no obvious consensus sequence for the

target sites.

The elements from Tripsacum have all the features of the maize sequences, and

members of the maize family are at least as diverged from each other as they are from

the Tripsacum elements. It seems reasonable to assume that all D sl sequences trace

back to a single element at some time in the past. We cannot make a distinction

between the alternatives of horizontal transfer of the elements, between maize,

Teosinte and Tripsacum or a vertical evolution from a common ancestor of the three

genera. If we make the assumption that the nucleotide substitution rate in these

elements is comparable to the neutral rates of nucleotide substitution which apply to

animal pseudogenes (5X 10-9 substitutions per site per year, Li, 1983) then the D sl elements duplicated and diverged from a common sequence between 8 and 25 million

years ago. It seems that these elements have been resident in the Maydeae genomes

for a long period of time.

D s l G E N E R A T E S A N A D D I T I O N A L I N T R O N I N T H E A d h l G E N E

Maize geneticists have reported a wide range of mutant phenotypes induced with

controlling elements and some of these effects are now understood at the molecular

level. An interesting example is the Adhl-Fm335 mutant in which the D sl element is

inserted into the transcribed leader sequence of the Adhl gene (Peacock et al. 1983;

Sutton et al. 1984). In a homozygous Adhl-Fm335 stock the amount of ADH1

enzyme is decreased to about 10% of the normal level but the specific activity

(Osterman & Schwartz, 1981) and temperature stability of the enzyme remain the

same (Sachs, unpublished results). How does the insertion of the D sl element in the

upstream region cause this mutant phenotype?

Northern hybridization analyses show that the length of the Adh 1 -specific mRNA

in the mutant is approximately the same as it is in the progenitor allele (Gerlach et al.1982); however, the amount of the mRNA is 100-fold lower. Under anaerobic

conditions the mutant plant shows a 20-fold increase in ADH1 enzyme activity and

an increase in Adhl -specific RNA of 20- to 50-fold. The kinetics of induction of Adhl mRNA levels are identical to the progenitor Adhl-F plant.

The transcribed regions of Adhl-Fm335 and AdhlF were compared directly in SI

mapping experiments using probes derived from the 5' region of the gene (Peacock

et al. 1984; Dennis el al. 1987). When the probe was prepared from the progenitor

allele, RNA from both the mutant and progenitor protected exactly the same length

fragment, but with a much weaker signal in the mutant. This indicates that all the

sequences present in the progenitor mRNA are also present in the Ds mutant mRNA

and that the transcription start site is exactly the same as in the progenitor. When the

mutant mRNA is used to protect a probe synthesized from the 5' region of the

mutant gene two fragments are seen. The first fragment extends 3' from the site of

insertion of the Ds element to the first exon-intron boundary, and the second

fragment, approximately 66 bp long, corresponds to a segment extending from the

start of transcription to a point 12 bases inside the D sl element. These results

indicate that only 12 bp of the Ds insertion are present in the mRNA of the mutant

gene and the remainder of the element is processed as an intron from the transcript

(Dennis et al. 1987). The intron donor sequence is 12bp from the D sl 5' terminus

and the intron acceptor site is at the junction between the 3' end of the D sl segment

and the Adhl leader sequence.

The processing of the D sl element from the mRNA does not in itself explain the

low level of messenger activity in the mutant. Run-on transcription experiments have

shown that the mutant has approximately the same rate of transcription as the

progenitor allele (L. Beach, personal communication), suggesting the low steady-

state level of mRNA is due to instability of the RNA in the mutant relative to the

progenitor allele.

128 W. J . Peacock and others

Ac/Ds transposable elements 129

The behaviour of the Ds element as an intron implies that sequences around the

donor and acceptor splice sites should resemble the consensus sequences seen for

splice sites in both plants and animals. They do, and moreover the sequence

TCCTAAC occurs 30bp before the 3' splice site. This sequence is identical to the

consensus lariat acceptor sequence (Ruskin et al. 1984) of introns and is located in

the correct position.

We conclude that the D sl element in the Adhl-Fm335 gene has all the necessary

sequence attributes to be spliced as an intron and that although transcription occurs

at a normal rate the RNA is less stable. This provides an example of how a

transposable element can introduce a new intron into a gene.

C R Y P T I C A c E L E M E N T S A N D T H E I R A C T I V A T I O N

One of the most perplexing problems with regard to transposable elements in

maize is their origin. Southern analyses with internal central probes of Ac support

the supposition that complete or near complete Ac segments may be present in maize

genomes, even though there is no detectable active Ac element (Fedoroff et al.1983). Other transposable element systems have been identified following genetic

trauma events. Peterson (1953) found the En (Spm) element in seed which had been

exposed to the Bikini atomic bomb blast and, more recently, Rhoades & Dempsey

(1982) described new transposable element systems in stocks exhibiting chromosome

breakage phenomena. It has been hypothesized that potentially active transposable

elements may be triggered into activity by conditions of genomic stress (McClintock,

1984).

An analysis which provides one possible explanation for the activation of cryptic

Ac elements has been carried out by Schwartz & Dennis (1986). They compared

three alleles of wx-m9; wx-m9 Ac, a Ds derivative (wxm9-Ds-cy) and revertants of

the derivative to an active Ac form (wx-m9 AcR). They cloned the waxy locus from

each derivative and showed that the Ac element was present in precisely the same

position in each and that there were no differences in restriction enzyme patterns of

the elements. However, when they examined the genomic organization of the

elements there were marked differences in restriction patterns with enzymes for

which the recognition sites are sensitive to methylation. The active Ac element is

methylated only at the left end, in contrast to the inactive (Ds) derivative which is

methylated at all H pall sites that occur throughout the elements. This particular Ds derivative reverts to an active Ac state, and in a number of revertants certain H pall sites at the right hand end of the element were no longer methylated. Different

revertants had different H pall sites unmethylated but all are grouped in the region

of the putative transposase gene promoter (see later). Although demethylation is

associated with reversion to the active Ac state, it is not complete and a number of the

H pall sites and all of the P vu ll sites remain methylated. It seems that particular

regions in the promoter segment are active or inactive depending upon their state of

130 W. 7- Peacock and others

méthylation. Dellaporta & Chomet (1985) described a change from an active to an

inactive state of the Ac element in the wx-m7 allele which was paralleled by a change

in digestibility of the P vu ll sites in the element. In this case, the méthylation state of

the H pall sites was not tested, nor were revertants isolated.

A N E F F E C T OF Ac O N E X P R E S S I O N OF A D E V E L O P M E N T A L L Y R E G U L A T E D

G E N E ?

One of the most interesting aspects of the effects of transposable elements on maize

gene expression is a phenomenon termed ‘presetting’ by McClintock. McClintock

(1964) observed the presetting phenomenon in her studies of two derivatives of the

a-m2 allele. In these a-m2 derivatives, the A gene, which is required for anthocyanin

pigmentation, is associated with the receptor of Spm ; when an active Spm is present

in the same nucleus, a variegated pigmentation pattern results. Interestingly,

McClintock recorded examples in which variegation could occur in the absence of an

active Sp m ; this variegation required the previous exposure of the a-m2 derivative

alleles to an active Spm element. McClintock concluded that the previous exposure

to the active Spm element had preset the a-m2 derivative alleles to produce a

variegated pigmentation pattern. The preset pattern was found infrequently, and it

was usually not heritable. However, examples of transmission of the preset pattern

through two generations were reported (McClintock, 1964, 1965).

McClintock (1967) reported other patterns of gene expression which she inter­

preted as presetting. In the c2-m2 allele, the C2 gene, which is required for plant and

kernel pigmentation, is controlled by the Spm element. Plants carrying c2-m2 produce ears with large sectors of coloured cob tissue indicative of early reversion of

c2-m2 to the active C2 allele. However, kernels within those sectors had aleurone

tissue with the variegated pattern of unchanged c2-m2. In this case, it seemed that

the Spm element at c2-m2 had preset the locus at an early stage in development to

produce certain patterns of expression later in development; a wild type (C2) pattern

in the cob tissue, and a variegated (c2-m2) pattern in the aleurone tissue.

Schwartz (1982) has proposed a similar presetting of the maize P locus involving

Ac. The P gene controls the accumulation of a red pigment produced in the pericarp

and cob tissues (Styles & Ceska, 1977). The pericarp, which is the outer layer of the

mature maize kernel, is formed by outgrowth of the ovary wall and hence is maternal

tissue. The dominant P-RR allele specifies red pigmentation in both pericarp and

cob, while the recessive P-WW allele conditions white pericarp and cob. Pigmen­

tation in cob and pericarp can be independently controlled according to the allelic

state of P. For example, the allele P-WR specifies white pericarp and red cob, while

P-RW determines red pericarp and white cob.

The P -W allele, which conditions variegated pericarp and cob, was shown by

Brink & Nilan (1952) to constitute an insertion of the transposable element Mp in the

P-RR gene. [Mp (Modulator) is functionally and molecularly equivalent to Ac and

Ac/Ds transposable elements 131

will be referred to as Ac from hereon.] As a consequence of its insertion in P -W , Ac suppresses P-RR expression. Transpositions which remove Ac from P -W restore P expression, resulting in red sectors of varying size on a colourless background.

Emerson (1917), and later Schwartz (1982), noted that on ears carrying P -W , pigmentation in two regions of the pericarp, the crown and the gown, could be

independently controlled. The crown comprises the region of silk attachment at the

top of the kernel, while the gown constitutes the sides of the kernel and that portion

of the top not including the crown. Ears carrying P -W often have large sectors of

kernels with pigmented crowns (dark crown kernels); the gowns of these kernels are

usually variegated, but may be fully red or white. Multi-kernel dark crown sectors are

invariably associated with a precisely coincident sector of pigmented cob. Although

these patterns of pigmented crown and cob occur in clonally-derived sectors, they are

not heritable (Emerson, 1917; Schwartz, 1982). These patterns resemble the

patterns of c2-m2 expression in cob and pericarp observed by McClintock (1967).

In order to explain the occurrence of non-heritable crown/cob pigmentation

patterns, Emerson (1917) proposed that the pericarp derives from two distinct cell

lineages which are set apart early in embryo development. One lineage of epidermal

origin gives rise to the pericarp crown and the floral parts of the cob. A second

subepidermal lineage gives rise to the pericarp gown and the megaspore mother cells.

Mutations of P -W to P-RR can occur independently in either cell lineage. In the

epidermal lineage, these mutations result in sectors of dark crown kernels over a

coloured cob sector. However, since this epidermal lineage does not contribute to the

megasporocytes, these mutations are somatic only and thus are not heritable.

Mutations of P -W to P-RR in cells of the subepidermal lineage produce sectors of

kernels with coloured pericarp gowns. Since this lineage also gives rise to the germ

cells, these mutational events are heritable. At Emerson’s suggestion, Randolph

(1926) performed a cytological study of the distribution of pigment in the pericarp

crown and gown in dark crown P -W kernels. However, Randolph found no

restriction of pigment to epidermal or subepidermal layers as predicted by Emerson’s

model.

In the light of the negative evidence reported by Randolph, Schwartz (1982)

proposed an alternative hypothesis for the non-heritable crown/cob pigmentation

patterns on P -W ears. Schwartz suggested that these patterns result from a

presetting of the P -W allele at an early stage of development to produce a particular

tissue-specific pattern of expression later in development.

We were encouraged by Schwartz to explore the molecular basis of the preset

patterns of P expression. Since the Ac element had been cloned previously (Fedoroff

et al. 1983), we used Ac as a probe to isolate a DNA clone derived from the P -W allele (Peterson & Schwartz, 1986). We found that the P -W clone contained an entire

Ac element with a restriction map identical to Ac elements previously isolated from

the maize waxy locus. Interestingly, the Ac at P -W was not bounded by short direct

repeats of maize genomic DNA. Identical results have been obtained by Lechelt

et al. (1986).

132 W. jf. Peacock and others

We then examined the structure of the P gene in P-RR revertants using as a probe

a sequence flanking the Ac element in the P - W clone. Southern blotting showed that

the Ac element had excised from the P locus in the P-RR revertant allele. This

finding confirmed our identification of the cloned P -W allele, and also indicated that

the 8 base pair flanking repeats are not necessary for excision of Ac.We initiated the presetting studies by examining a P allele derived from kernels

within a sector of mutant tissue on a P-W /P -W R ear. Kernels within this sector had

crowns which were completely pigmented, lightly pigmented, or unpigmented. The

gowns of the kernels within this sector were largely unpigmented, in marked contrast

to the variegated gowns on kernels outside the mutant sector. Any contribution of the

P -W allele to cob pigmentation was masked by expression of the P-WR allele which

specifies red cob. Plants grown from kernels within the mutant sector produced ears

with completely colourless pericarp and cob; the crown pigmentation pattern was not

inherited. Further genetic analyses showed that the progenitor P -W allele had

mutated to a P-WW allele and had lost the Ac element.

We examined the molecular structure of the derivative P-WW allele using as probe

a DN A fragment located 3 kb from the Ac element in the P -W allele. This probe

detects RNA transcripts found specifically in the maize pericarp and cob tissue, and

not in the embryo and endosperm, suggesting that the probe represents a part of the

P structural gene. In Southern analyses, comparing the progenitor P -W and the

derivative P-WW allele we found that the P-WW allele had a deletion which removed

sequences homologous to the probe. The deletion may have resulted from imprecise

excision of Ac. These results show that the embryos (and presumably the pericarp

gown) of the kernels within the original mutant sector carried the P-WW deletion

allele. We suggest that the pericarp crown tissue carried a P -W allele, and that,

within this crown tissue, reversions to P-RR by excision of Ac produced the dark

‘crown pigmentation seen in the original mutant sector.

We suggest that these data fit the hypothesis originally proposed by Emerson

(1917) and supported recently by Greenblatt (1985). An even more important

question is whether the crown/ cob and gown tissues develop from the same cell

lineage, or from two lineages separated early in development. In our experiments, if

the gown and crown cells were derived from the same ancestral cell which suffered

the P gene deletion, we would expect that the sector should consist of completely

white kernels. The occurrence of kernels with pigmented crowns within the P-WW sector is consistent with the notion that the crown/cob and gown tissues are derived

from different cell lineages, since a deletion of the P gene occurring within the gown

lineage would not affect P gene expression in the crown/cob lineage. Clearly,

additional data are required to test this inference. However, if valid, this cell-lineage

explanation of the non-heritable dark crown pigmentation pattern might also apply to

the previously reported patterns of expression of the P -W and c2-m2 alleles which

were attributed to genetic presetting. These data do not bear on the original

observations of presetting at the a-m2 allele, but they emphasize the interaction of

transposable element, gene and development in determining a particular phenotype.

Ac/Ds transposable elements 133

A c AS A G E N E T I C T O O L I N H E T E R O L O G O U S S P E C I E S

There is evidence that Ac can catalyse its own transposition in a heterologous

species by a mechanism of excision that is probably identical to that used in maize

(Baker et al. 1986). An Ac element (Ac wx-m9) with short flanking segments of waxy DN A was introduced into tobacco cells via a T-DNA vector. In almost half of the

transformed calli analysed, Ac had transposed, creating an ‘empty site’ within the wx DN A segment. Sequencing showed that the excision event was of the type observed

in Ac/Ds revertants in maize, i.e. with small deletions and/or base changes adjacent

to the site of the original insertion. Parallel experiments using a Ds element derived

from Ac wx-m9 (Ds wx-m9) did not result in excision or transposition of Ds, implying that Ac encoded function(s) were responsible for the transpositions seen in

tobacco cells.

In maize, Ac has already been used to isolate genes in which an Ac insertion has

resulted in an altered phenotype (Fedoroff et al. 1984). In this procedure, the gene

of interest is ‘tagged’ by an Ac or Ds insertion, then identified by hybridization to an

Ac probe. One problem with this approach is the difficulty of associating changes in

mobility of any Ac hybridizing band with the mutation because of the high number

of copies of Ac and its derivatives in the maize genome. In a heterologous system,

where there is no background of Ac related sequences, transposon tagging of genes

and the correlation of a mutant phenotype with an alteration in the pattern of bands

hybridizing to Ac may be simplified. The mutant allele could then be used to clone

the corresponding wild type gene.

The ability of transposable elements to jump from one region of DNA to another

suggests that they might be harnessed for use as transformation vectors. Such a

system has already been developed using P elements in Drosophila (Spradling &

Rubin, 1982). The demonstration of normal excision and integration of Ac in the

tobacco genome indicates that Ac could potentially be of value as a vector to facilitate

the integration of defined DNA segments into a range of plant genomes. In its

simplest form, the gene to be transferred would be placed with a selectable marker

within the Ac element. Following transfer of the engineered Ac into plant cells,

either by a direct transfer method or via a Ti plasmid vector, the transposase

function of the Ac element would excise the Ac segment from its carrier plasmid and

reinsert it into host plant DNA. A complication of this approach is that the gene to be

transferred must not be integrated into the Ac element at a site which disrupts the

transposase.

A second approach which avoids this problem is to place the DNA to be

transferred within an Ac or Ds element and to rely on an intact Ac element cloned on

a separate plasmid to produce the transposase activity in trans. We have investigated

the feasibility of this approach for the stable transformation of tobacco protoplasts.

When a kanamycin resistance marker inserted into the D sl transposable element was

electroporated into tobacco protoplasts with or without Ac provided in trans, no

substantial Ac-dependent increase in the number of kanamycin resistant colonies was

134 W. J . Peacock and others

obtained. This result suggests that additional modifications to the system may be

necessary in order to produce useful levels of transformation enhancement.

One modification which may be effective is the replacement of weak endogenous

Ac promoters with much stronger promoters, such as the Cauliflower Mosaic virus

(CaMV) 35S promoter. A prime candidate for such a replacement is the region

upstream of the 167/168 codon ORF (Fig. 1), since this segment appears to have

weak promoter activity and is just upstream of a transcribed region (see next section).

We have replaced this region with the CaMV 35S promoter and have detected

enhanced transcription of the downstream region, however data are not yet available

regarding the effect of this modification on functional activity.

The tobacco system also offers the prospect of being able to analyse the function of

Ac by in vitro manipulation of Ac, followed by the introduction of the mutated Ac and subsequent analysis of its expression and function.

T R A N S C R I P T I O N OF Ac

Kunze & Starlinger (1986) have reported three different size classes of poly A

RNA that hybridized to an Ac probe in Northern analysis. Only the largest transcript

(3-5 kb) appears to be Ac specific - the smaller transcripts may be derived from

transcription of Ds elements inserted into transcriptionally active DNA, since these

are observed in both Ac containing and Ac free plants. Kunze and Starlinger suggest

that the single large transcript from Ac is initiated about 150 bp from one inverted

repeat and terminates at a position 264 bp from the other end, although cDNA clones

isolated thus far lack the 5' terminal 500 bp. Comparison of these cDNA clones with

the genomic Ac sequence indicates the presence of four introns comprising 650 bases

in the primary transcript.

Computer analysis of the sequence of the Ac element (Pohlman et al. 1984a,b) shows that in one orientation there are nine open reading frames in excess of 75

codons, which together encompass 3-5 kilobases of DNA; in the opposite orientation

there are five such ORFs, encompassing only 1-3kb of DNA (Fig. 1). The ORFs

observed, if transcribed, could represent complete coding regions of proteins, or

could be spliced to form larger ORFs by the removal of introns. We therefore

directed our initial efforts to obtaining evidence for Ac transcription of the 5' region

of the strand carrying the longest ORFs.

Two approaches were taken in the analysis of this region. Segments of Ac DNA

located upstream of both the 167/168 codon ORF and the 485 codon ORF were

positioned upstream of a chloramphenicol acetyltransferase gene in a CAT ex­

pression vector and transferred into Nicotiana plumbaginifolia protoplasts by

electroporation. The vector also carried a region of the octopine synthase gene which

enhances the expression of known promoter segments (Ellis et al. 1987). After 24 h

incubation, the protoplasts were assayed for CAT activity. Low levels of promoter

orientation-dependent CAT activity were detected for both the 167/168 upstream

region and the 485 codon upstream region.

Ac/Ds transposable elements 135

The second approach was to determine whether the 167/168 codon ORF and the

485 codon ORF regions are transcribed. Probe DNA complementary to the possible

transcripts associated with these regions was prepared and used in SI mapping of

total RNA and poly A RNA isolated from Ac+ or Ac~ seedlings. The seedlings were

grown from seeds, taken from several independent cobs, which had been separated

into Ac+ and Ac~ on the basis of variegation in aleurone colour. As shown in Fig. 2,

probes la and lb initiate at the same point 693 bases 3' of the BamW\ site and differ in

length by 110 bases (800 bases and 690 bases, respectively).

Two protected fragments were observed when either one of these probes was used.

The sizes of the protected fragments were estimated to be 510 and 570 bases in length

and are identical for both probes, indicating that the region between theBamHI and

Pvul sites is not transcribed. The intensity of the two bands was approximately

102

294

84

<==]

485

105

C = 3168

< ..... " I167

128

Probe lb

Probe 2 Probe la

Fig. 2. Map of Ac transposable element showing location of probes used in Si analysis.

Restriction site coordinates and open reading frames (see also Fig. 1) were determined

from sequence data of Pohlman et al. (1984a,b). Probe DN A was prepared by primer

extension of segments of Ac cloned into single strand vectors (Yanisch-Perron et al. 1985). Probes la and lb were initiated from a 19 base primer and digested with either BamHl or

Pvul . Probe 2 was initiated from a site on the vector adjacent to the Hind i l l site and

digested w ithP tu ill. Uniformly labelled probes were hybridized to RNA, treated with Si

nuclease, and analysed on polyacrylamide sequencing gels.

136 W. J . Peacock and others

equal and they were much more intense in RNA derived from Ac-containing

seedlings. The same bands were seen when either total or poly A RNA was used to

protect the radioactive probe. The appearance of these bands in Ac~ material may

arise from a small number of Ac+ seeds due to misclassification of the Ac-containing

seed. In addition, the possibility that the transcripts observed arise from Ac-derived

Ds elements present in the maize genome cannot be excluded.

The shorter fragment is consistent with an intron splice junction located at

nucleotide 4226, approximately 500 bp from the probe initiation site. A consensus

intron acceptor site is located at this position (Mount, 1982). Other intron-like

characteristics of the upstream region include a potential lariat sequence (Ruskin

et al. 1984) located at nucleotides 4292-4299 and three potential donor splice sites

located at 4351-4361. Translation could begin at the ATG codon located a short

distance upstream. Alternate initiation, splicing, or termination, could account for

the 570 base protected fragment.

Probe 2 (Fig. 2) covers the region from Hind.Hl to P v u ll, a distance of 468 bases,

all of which lies within the 485 codon ORF. A fragment of approximately 470 bases is

protected by poly A RNA isolated from Aocontaining seedlings. A 194 bp deletion 3'

of this H in d lll site within the ORF has been shown to block autonomous

transposition in maize (Fedoroff et al. 1983). Together these results suggest that the

485 codon ORF is transcribed and translated into part or all of a protein essential for

transposition.

It is anticipated that nuclease Sj mapping with different probes should enable us to

define more exactly the regions of Ac that are represented as RNA. Northern analysis

should provide information on the size and number of transcripts while cDNA

cloning and sequencing will precisely define the intron/exon boundaries. Our

interest lies in the mechanism of Ac/Ds transposition and in the nature of protein­

DNA interaction in the transposition and expression of Ac. One approach to this

problem is the generation of Ac encoded proteins in vitro for use in DNA binding

studies.

C O N C L U D I N G R E M A R K S

Although we now have a molecular description of several Ac/Ds transposable

elements and the events accompanying their insertion and excision, little is known

about their movement and how it is regulated. Our approach to this question will be

to try to characterize the RNA and protein products of these elements, as well as the

regulatory sequences which control their expression. In this way we hope to address

some of the remaining questions about Ac activity, e.g. whether there is developmen­

tal control of Ac movement, whether genomic stress causes Ac activation, how

methylation/demethylation might function to regulate Ac movement, and what is

the basis for the inverse dosage effect of Ac.The study of transposable elements at the genetic and molecular levels has already

yielded valuable insight into how these elements function and interact with their

target genes. We anticipate that continued study of these elements at the molecular

Ac/Ds transposable elements 137

level will be of practical benefit in exploiting these elements for gene tagging and

mutagenesis, and as potential vectors for gene transfer.

R E F E R E N C E S

Baker, B ., S c h e l l , J., L ö r z , H . & F e d o ro ff , N . (1986). Transposition of the maize controlling

element ‘Activator’ in tobacco. Proc. natn. Acad. Sei. U.S.A. 83, 4844-4848.

Behrens, U ., F e d o ro ff , N ., L a ird , A ., M ü lle r- N e um ann , M ., S ta r l in g e r , P. & Y ode r, J.

(1984). Cloning of the Zea mays controlling element Ac from the w x-m l allele. Mol. Gen. Genet. 194, 346-347.

B rink , R. A. & N ila n , R. A. (1952). The relation between light variegated and medium variegated

pericarp in maize. Genetics 37, 519-544.

Courage-Tebbe, U., D ö r in g , H.-P., F e d o ro ff , N. & S ta r lin g e r , P. (1983). The controlling

element Ds at the shrunken locus in Zea m ays: structure of the unstable sh-m5933 allele and

several revertants. Cell 34, 383.

D ellaporta , J. C. & Chom et , P. S. (1985). The action of maize controlling elements. In Genetic Flux in Plants (ed. B. Hohn & E. S. Dennis), pp. 169-216. Vienna: Springer-Verlag.

D enn is , E. S ., G e r la c h , W. L ., Peacock, W. J. & S ch w artz , D . (1986). Excision of the Ds controlling element from tht Adhl gene of maize. Maydica 31, 47-57.

D e n n is , E. S., Sac h s , M. M ., G erlach , W. L ., Bea ch , L. & Peacock , W. J. (1987). TheDs/

controlling element acts as an intron in the Adhl mutant Fm335 (unpublished).

D ö r in g , H. P., T il lm a n n , E. & S ta r lin g e r , P. (1984). D N A sequence of the maize transposable

element Dissociation. Nature, Lond. 307, 127-130.

E l l is , J. G ., L le w e l ly n , D . J., D enn is , E. S. & Peacock, W . J. (1987). Maize Adhl promoter

sequences control anaerobic regulation: addition of upstream promoter elements from constitut­

ive genes is necessary for expression in tobacco. EMBO J. 6, 11-16.

E m erson , R. A. (1917). Genetical studies of variegated pericarp in maize. Genetics 2, 1-35.

E ng e ls , W. R. & Preston , G. R. (1981). Identifying P factors in Drosophila by means of

chromosome breakage hotspots. Cell 26, 421-428.

F e d o ro ff , N ., F u rte k , D . & N e lson , O. (1984). Cloning of the bronze locus in maize by a simple

and generalizable procedure using the transposable controlling element Activator (Ac). Proc. natn. Acad. Sei. U.S.A. 81, 3825-3829.

F ed oroff , N ., W essler , S. & Sh u r e , M . (1983). Isolation of the transposable maize controlling

elements Ac and-Ds. Cell 35, 235-242.

F re e lin g , M. (1984). Plant transposable elements and insertion sequences. A. Rev. PI. Physiol. 35, 277-278.

G er la c h , W . L ., D e n n is , E. S ., Peacock , W . J. & C leg g , M . T . (1987). The D sl controlling

element family in maize and its grass relative Tripsacum. J . mol. Evol. (in' press).

G e r la c h , W . L ., P ryor, A. J., D enn is , E. S., F e r l, R . J., Sachs, M . M . & Peacock, W . J.

(1982). cD N A cloning and induction of the alcohol dehydrogenase gene (Adhl) of maize. Proc. natn. Acad. Sei. U .SA . 79, 2981-2985.

G reenblatt , I. (1985). The pericarp of maize: a major tool to study transposable elements. In

Plant Genetics, U C LA Symposia on Molecular and Cellular Biology 35, 405-418.

K u n ze , R. & S ta r lin g e r , P. (1986). Studies of Ac-derived RNA. M aize Newsletter 60, 39.

L e c h e lt , C., L a ird , A. & S ta r lin g e r , P. (1986). Cloning of DN A from the P locus. Maize genetics Coop. Newsletter 60, 40.

L i, W.-H. (1983). Evolution of duplicate genes and pseudo genes. In Evolution of Genes and Proteins (ed. M . Nei & R. K. Koehn), pp. 14-37. Sunderland, MA: Sinauer Associates.

M c C lin to c k , B. (1950). Mutable loci in maize. Carnegie Inst. Wash. Year Book 49, 157.

M c C lin to c k , B. (1951). Chromosome organization and genic expression. Cold Spring Harbor Symp. Quant. Biol. XVI, 13-47.

M c C lin to c k , B. (1956). Controlling elements and the gene. Cold Spring Harbor Symp. Quant. Biol. XXI, 197-216.

M c C lin to c k , B. (1964). Aspects of gene regulation in maize. Carnegie Inst. Washington Year Book 63, 592-602.

138 W. J . Peacock and others

M c C lin to c k , B. (1965). Components of action of the regulators Spm and Ac. Carnegie Inst. Washington Year Book 64, 527-534.

M c C lin to c k , B. (1967). Regulation of pattern of gene expression by controlling elements in

maize. Carnegie Inst. Washington Year Book 65, 568-578.

M cC l intock , B. (1984). The significance of responses of the genome to challenge. Science 226,

792-801.

M ikus , M . D . & Petes, T. D . (1982). Recombination between genes located on non homologous

chromosomes in Saccharomyces cerevisiae. Genetics 101, 369—404.

M o u n t, S. M . (1982). A catalogue of splice junction sequences. Nucleic Acids Res. 10, 459-472.

M U lle r-N eum ann , M ., Y oder, J. I. & S ta r l in g e r , P. (1984). The DN A sequence of the

transposable element Ac of Zea mays L. Mol. Gen. Genet. 198, 19-24.

OsTERMAN, J. L. & S c h w a r t z , D . (1981). Analysis of a controlling-element mutation at the Adh locus of maize. Genetics 99, 267.

Peacock, W. J., D enn is , E. S., G e r la c h , W. L ., L le w e lly n , D ., LO rz, H., Pryor, A. D .,

Sachs, M. M ., S ch w artz , D . & S u tto n , W. D . (1983). Gene transfer in maize: controlling

elements and the alcohol dehydrogenase genes. In Proceedings of the Miami Winter Symposia (ed. K . Downing et al.), 20, p. 311. New York: Academic Press.

Peacock, W . J., D enn is , E . S., G e r la c h , W . L ., Sachs, M. M. & S ch w artz , D . (1984).

Insertion and excision of Ds controlling elements in maize. In Recombination at the DNA Level, Cold Spring Harb. Symp. Quant. Biol. 49, 347—354.

P eterson , P. A. (1953). A mutable pale green locus in maize. Genetics 38, 682.

Peterson , T. & S ch w artz , D. (1986). Isolation of a candidate clone of the maize P locus. M aize Genetic Coop. Newsletter 60, 36-37.

P oh lm an , R. F ., F e d o ro ff , N . V. & Messing, J. (1984a). The nucleotide sequence of the maize

controlling element Activator. Cell 37, 635-643.

P oh lm an , R. F ., F e d o ro ff , N. V. & M essing, J. (19846). Correction: Nucleotide sequence of

Ac. Cell 39, 417.

R a n d o lp h , F. R. (1926). A cytological study of two types of variegated pericarp in maize. Memoir 102. Cornell University Agricultural Experiment Station, Ithaca, N.Y.

R a n d o lp h , L. F. (1936). Developmental morphology of the caryopsis in maize. J . Agric. Res. (Washington, D .C .) 53, 881-916.

R h o a d e s , M. M. & D empsey , E. (1982). The induction of mutable systems in plants with the

high-loss mechanism. Maize Genetics Coop. Newsletter 56, 21—26.

R usk in , B., K ra in e r , A . R ., M an ia tis , T. & G reen , M . R . (1984). Excision of an intact intron as

a novel lariat structure during pre-mRNA splicing in vitro. Cell 38, 317-331.

Sachs, M. M ., Peacock, W. J., D enn is , E. S. & G e r la c h , W. L. (1983). Maize Ac-Ds controlling elements - a molecular viewpoint. Maydica 28, 289.

S ae d le r , H. & Nevers, P. (1985). Transposition in plants: a molecular model. E M B O J. 4, 585-590.

Sc h w a r t z , D . (1982). Tissue-specific regulation of gene function: Presetting and erasure. Proc. natn. Acad. Sci. U.S.A. 79, 5991-5992.

S ch w artz , D . & D e nn is , E. S. (1986). Transposase activity of theAc controlling element in maize

is regulated by its degree of methylation. Mol. Gen. Genet. 205, 476-482.

S p rad lin g , A. C. & R ub in , G. M. (1982). Transposition of cloned P elements into Drosophila germ line chromosomes. Science 218, 341.

S ty les , E. D . & Ceska, O. (1977). The genetic control of flavonoid synthesis in maize. Can. J . Genet. Cytol. 19, 289-302.

S u tto n , W. D ., G e r la c h , W. L ., S ch w a rtz , D . & Peacock, W. J. (1984). Molecular analysis of

Ds controlling element mutations at theA dh l locus of maize. Science 223, 1265-1268.

W eck, E ., C ourage , U ., D O ring , H.-P., F ed o ro ff , N. & S ta r lin g e r , P. (1984). Analysis of sh- m6233, a mutation induced by the transposable element Ds in the sucrose synthase of Zea mays. EMBOy. 3, 1713.

W essle r , S. R., B aran , G ., V a ragona , M . & D e lla p o r ta , S. L. (1986). Excision of Ds produces waxy proteins with a range of enzymatic activities. EM BO J. 5, 2427-2432.

Yanisch-Perron, C., V iera , J. & M essing, J. (1985). Improved M13 cloning vectors and host

strains: nucleotide sequences of the M13mpl8 and pUC19 vectors. Gene 33, 103-119.