Plant Molecular Biology 26:1521-1528, 1994. © 1994 Kluwer Academic Publishers. Printed in Belgium.

1521

Activation tagging: a means of isolating genes implicated as playing a role in plant growth and development Richard Walden Jeff Schell 1

1,,,

Klaus Fritze

1, Hiroaki Hayashi 1,3,

Edvins Miklashevichs 2, Hinrich Harling I and

1Max Planck Institut ffir Zuchtungsforschung, Carl yon Linnd Weg 10, D-50829 KOln, Germany (*author for correspondence)," 2Institute of Microbiology, Kleisti, L V 1067 Riga, Latvia; 3Current address." Faculty of Agriculture, University of Tokyo, Bunkyo-ku, Tokyo 113, Japan Received 21 April 1994; accepted 26 April 1994

Key words: activation tagging, Arabidopsis, auxin, cytokinin, polyamines, signal transduction, T-DNA tagging

Abstract

Activation T - D N A tagging has been used to generate a variety of tobacco cell lines selected by their ability to grow either in the absence of auxin or cytokinin in the culture media, or under selective levels of an inhibitor of polyamine biosynthesis. The majority of the cell lines studied in detail contain single T-DNA inserts genetically co-segregating with the selected phenotype. While most of the plants regenerated from the mutant cell lines appear phenotypically normal, several display phenotypes which could be inferred to result from disturbances in the content, or the metabolism, of auxins and cytokinins, or polyamines. The tagging vector is designed to allow the isolation of tagged plant genes by plasmid rescue. Confirmation that the genomic sequence responsible for the selected phenotype has indeed been isolated is provided by PEG-mediated protoplast D N A uptake of rescued plasmids followed by selection for protoplast growth under the original selective conditions. Several plasmids have been rescued from the mutant lines which confer on transfected protoplasts the ability to grow either in the absence of auxin or cytokinin in the culture media, or under selective levels of an inhibitor of polyamine biosynthesis. This review describes the background to activation tagging and our progress in characterizing the genes that have been tagged in the mutant lines we have generated.

Introduction

The power of mutant analysis in plants as a means of dissecting complex biochemical and developmental processes has only recently been fully realized by molecular biologists. While the biochemical analysis of mutant plant lines has provided valuable insight into the consequences of the mutation, it is now becoming feasible to isolate and characterize the mutated genes them-

selves, hence allowing an understanding of the molecular basis of the mutation itself [4, 13, 11]. Transposon and T-DNA tagging, as well as chromosome walking, have all proved to be valuable means of isolating mutated plant genes [7]. In particular, T-DNA tagging in Arabidopsis has proven to be an especially powerful means ofgene isolation and an ever increasing number of genes tagged in this manner are being described [6, 15]. Central to each of these methods of mutant [285]

1522 generation is the practical requirement that mutations, which generally involve disruption of a transcriptional unit, result routinely in a recessive mutation. This means that the mutant phenotype can only be visualised following selflng of the mutated individuals and this demands not an insubstantial amount of effort. Moreover, the generation of a specific mutation is fortuitous. We have developed an alternative method of T-DNA tagging where the mutation is a consequence not of gene disruption, rather the activation of expression of the tagged gene. This has come to be known as activation tagging [30, 31 ]. Because the mutation itself results from the activation of gene expression the resulting phenotype is dominant. This allows a direct selection for a desired phenotype from amongst the population of primary transformants. To achieve the deregulation of expression of flanking plant genes upon insertion into the plant genome we have engineered a T-DNA tag to contain the transcriptional enhancer sequence of the cauliflower mosaic virus (CaMV) 35S RNA promoter cloned as a tandem tetramer near the border of the T-DNA. Mutant cell lines are generated by Agrobacterium protoplast co-cultivation. Thus a large number of mutant individuals can be generated with ease and a chemical selection can be applied to the primary transformants. Mutant cell lines generated in this manner can be cultured further and ultimately, plants regenerated. We have used activation tagging to generate mutants affected in processes which have been previously proved to be relatively difficult to mount a genetic analysis, namely, the action of the plant growth substances auxin and cytokinin, as well as the polyamines. The notion underlying these experiments is that growth factors are likely to exert their effects through the induction ofgene expression. By the activation tagging of a gene which normally becomes active only in the presence of a particular growth factor and whose action results in cell growth and division, we might isolate genes that play a role in a specific signal transduction pathway. In the case of isolating genes whose overexpression results in callus growth in the absence of growth substances we [286]

are, in effect, isolating plant cellular protooncogenes.

Activation tagging The development of activation tagging exploits several experimental observations. First, protoplast Agrobacterium co-cultivation is an effective means of producing extremely large numbers of transgenic plant cells with relative ease [3, 23]. Second, during the transformation process the T-DNA of Agrobacterium tumefaciens inserts preferentially into potentially transcribed regions of the plant genome [ 16]. Third, there is a multiplying effect on activity when transcriptional enhancer sequences derived from a promoter sequence are cloned as multimers and they can activate gene expression distant from their site of insertion [ 12]. Finally, central to the use of activation tagging is the notion that the majority of mutations produced by overexpression of a gene will be dominant which in turn allows direct selection for a specific phenotype from amongst the population of primary transformants. The tagging vector we have developed is based on the pPCV series of plant transformation plasmids [ 17]. The plasmid backbone contains origin of replication and mobilisation sequences allowing maintenance in Agrobacterium and left (LB) and right (RB) border sequences delimiting the T-DNA containing a hygromycin resistance gene at the LB sequence driven by the nopaline synthase promoter, the high-copy-number Escherichia coli plasmid pIC19H [22], and located 20 bp from the RB sequence the transcriptional enhancer sequence (-90 to -420) [27] of the 35S RNA promoter of CaMV cloned as a tetramer (Fig. 1). The idea of using such a T-DNA as a tag is that following insertion into the plant genome the multiple enhancers will activate the expression of flanking plant genes and that the plasmid allows rescue of flanking sequences with ease. The production of a dominant mutation has a further advantage in that functional testing of flanking plant sequences can be carried out by transfecting protoplasts with the rescued D N A

1523 r

ori g4pA

Amp

HPT

pnos L,B.

R.B.

ori O

ori T

Fig. 1. Structure of the activation tagging vector pPCVICEn4HPT. Functional regions of the plasmid are shown, oriV and oriT are responsible for stable maintenance in Agrobacterium while between the left (L.B.) and right (R.B.) borders of the T-DNA is the hygromycin resistance gene (HPT) fused with the nopaline synthase promoter and the poly A sequences ofgene 4 of the Agrobacterium T-DNA, an ampicillin resistance gene (Amp) and origin of replication (ori) functional in E. coli and the enhancer sequence of the 35S RNA promoter cloned in tandem (arrows).

and selecting for growth under the original selection pressure. This then provides the ultimate proof that the tagged gene responsible for the selected phenotype has indeed been isolated and opens the way for its localization by further testing of deletion derivatives of the rescued plasmid. We have used tobacco protoplast Agrobacterium co-cultivation to generate mutant cell lines transformed with the activation tag which have the ability to grow in tissue culture in the absence of either auxin or cytokinin in the culture media, or in the presence of MGBG, an inhibitor of polyamine biosynthesis. By this means we aim to create novel mutants modified in the metabolism or signal transduction of these plant growth substances, or polyamines.

Auxin-independent mutants Although auxin has been long recognized as playing a central role in plant growth and development, the molecular basis of its action is little understood [25, 26]. While several experimental systems are available to study auxin action

[2, 10, 29], debate remains concerning how auxin modulates its differing effects, such as cell elongation, cell division, side shoot growth and lateral root initiation and growth, either upon external application, or through the expression of transgenes involved in auxin biosynthesis [1,14, 28]. There is, however, general agreement that auxin is required in the culture media for protoplasts to divide and form callus [18, 24, 25]. We decided to exploit this in generating tagged mutants that might be modified in auxin biosynthesis, perception, or signal transduction. With this in mind, following tobacco protoplast tranformation with agrobacteria containing the activation gene tag, we selected for growth of transgenic cells in the absence of auxin in the culture media. In a typical experiment, using 30 × 106 protoplasts and with a transformation frequency of about 2030~o, we obtained 13 calli growing under these conditions. The majority of these calli regenerated into plants. Some of the regenerated plants display phenotypic changes such as increased frequency of side shoot formation, reduction of root initiation and growth as well as premature senescence, all of which could be considered as indicative of changes in auxin metabolism and action. The majority of regenerated plants however displayed no obvious phenotypic changes. Nevertheless, in all examples studied to date, protoplasts reisolated from the plant lines divide in culture in the absence of auxin in the media, and this characteristic genetically co-segregates with a single T-DNA insert. By Southern analysis, we judge that at least eight differing genes have been tagged in separate individuals. We have studied one plant line, axi 159 in most detail ([9], Walden et aL, submitted). In axi 159 the T-DNA has inserted ca. 6000 bp downstream from the transcriptional start site of the gene responsible for auxin-independent growth, axi 1. axi 1 itself is a 4111 bp gene split by nine introns and apparently is a member of a small multigene family in tobacco. Northern analysis indicates that in untransformed tobacco plants axi I transcripts accumulate to highest levels in root tissue. In the tagged plant line, axi I transcripts are detectable in all tissue tested, confirming the notion [287]

1524 of the experiment that upon insertion of the tag into the plant gene the expression of axi i has become deregulated. In freshly isolated protoplasts from untransformed plants axi I expression requires auxin in the culture media. In protoplasts isolated from a x i l plants, a x i l transcripts accumulate both in the presence and absence of auxin in the culture media. The coding sequence of the gene displays no obvious similarities to genes that have been previously characterized. Nevertheless, preliminary work suggests that axi I expression has an effect on gene expression in both transient assays and in intact plants (Walden, Cjaza and Bongartz, unpublished). Naturally our interest currently is focused on determining how direct, or indirect, this effect might be. Similarly, a number of other genes have been isolated from tobacco lines that are able to promote auxin-independent growth (Harling, Miklashevichs and Walden, unpublished).

Cytokinin independent mutants Similar to the experiments described above we have generated tagged cell lines that have the ability to grow in the culture media in the absence of cytokinin. One of these lines regenerates plants that are dwarfed only with difficulty. The others, although able to regenerate, are all male- and female-sterile. This obviously precludes genetic analysis but in two examples plant D N A flanking the T-DNA insert has been rescued and, intriguingly, these are able to promote growth in the absence of both auxin and cytokinin (Miklashevichs and Walden, unpublished).

Mutants modified in polyamine metabolism In the previous examples selection was based on the generation of mutants able to grow in the absence of either auxin, or cytokinin, in the culture media. However, with such a tagging system dominant selection, i.e. growth under normally toxic levels of a particular compound, is also fea-

[2881

sible. To test this, and to establish a means of generating mutations in polyamine metabolism or perception, we have used the system to generate cell lines selected to grow in culture in the presence of selective levels of MGBG, an inhibitor of polyamine synthesis. The aim here was to generate mutants in polyamine metabolism, so that the role of polyamines in plant growth and development, an area that remains a matter of debate, might be assessed [5]. Methylglyoxal-bis(guanylhydrazone) (MGBG), is an inhibitor of S-adenosylmethionine decarboxylase (SAMdc), a key enzyme in the polyamine biosynthetic pathway being the ratelimiting step in the conversion of putrescine to spermidine and spermine. Previously, Malmberg and coworkers have generated through UV mutagenesis a variety of tobacco lines resistant to MBGB [20]. Plants regenerated from these mutant cell lines displayed a variety of phenotypes including aberrant flower formation with altered patterns of morphogenesis including organ replacement and sexual inversion. In some cases the biochemical basis of the mutation were partially characterized [21 ]. However, the severity of the phenotypes displayed by the regenerated plants precluded detailed genetic analysis, though in two examples the phenotypes observed appeared to segregate as a single dominant genetic locus [21, 19]. Building on these observations, we decided to apply the strategy of T-DNA activation tagging in an attempt not only to generate mutants modified in polyamine metabolism, but also to isolate the genes involved. We carried out activation tagging and recovered eight independent mutant lines from a tagging experiment involving 30 × 10 6 protoplasts. All lines regenerated into plants without apparent difficulty. Two of the lines displayed dramatic phenotypic changes including reduced internode length and leaf twisting. However, the most characteristic changes were seen in flower morphology where malformed flowers were malesterile and possessed elongated styles. In one line parthenocarpy was frequently observed in unpollinated flowers. Two of the plant lines studied in detail contained increased levels of SAMdc ac-

1525 I

a x i 15g

I

I kb

E

K .

I

/

~-.,.-.q

K

,

E:.-..<::]

illillllllill

K

c

B

KE

I

I

I

I I

4/1D K

B

g

K

B K 1111111111111~1]]

Ki~:::::::I

1811

El

~\',,i

K i:i~@':.]

[k'-,,.'~

El

i/e

I

I

!:11111111

I

Fig. 2. Organization of the activation T - D N A tag in the genome of differing mutant lines. The organization of the T - D N A in three lines selected for growth in the absence of auxin (axi159, 4/1D and 10D), or in the absence of cytokinin (1/0) in the culture media are shown. The borders of the T - D N A are indicated by the inverted arrows and the internal organisation of the T - D N A is as in Fig. 1. In two cases the T - D N A has inserted as a complete (4/1D) or partial (10D) inverted dimer. The position of the region of the plant D N A demonstrated to direct either auxin, or auxin and cytokinin independent growth following plasmid rescue is indicated in each case by the broken line. Restriction enzyme sites are: Bam HI (B), Cla I (C), Eco RI (E) and Kpn I (K).

tivity compared with untransformed tobacco. In one plant line this increase in enzymatic activity is coupled with an approximate doubling in the level of putrescine and also significantly higher levels of spermidine compared with untransformed plants. These increases in free polyamines are accompanied by increases also in conjugated polyamines. Intriguingly, in the other plant line the levels of both free and conjugated putrescine, as well as free spermidine, are approximately the same as in untransformed callus but the levels of conjugated spermidine are reduced by about 50~o. Plasmid rescue was used to recover the T - D N A and flanking plant sequences from one of the mutant lines and this, when transfected into protoplasts, confers MGBG-resistant growth (Fritze, Cjaza and Walden, submitted). Northern analysis indicates that the rescued plant D N A is overexpressed in the mutant plant line when compared with wild type tobacco and work is currently in progress to clone c D N A s corresponding to the rescued genomic sequences.

Discussion

One of the most direct ways of dissecting a complex biological process is mutant generation and analysis. Gene tagging provides a means not only of generating mutations but also of isolating the genes involved in a specific process. Routinely, mutants resulting from tagging, be it transposon or T - D N A tagging, are recessive making recovery of specific mutants fortuitous. Activation tagging, through the generation of dominant mutations, allows a means of selecting for a specific mutation. The selection schemes we have used, growth in the absence of a necessary growth substance, or in the presence of a selective chemical, relies on little prior knowledge of the process under study. Nevertheless, the mutations that are generated have one characteristic in common: the expression of the tagged gene allows growth under selective conditions. Thus, mutations that we recover could result from a variety of events including: the activation of a specific developmental [2891

1526 pathway, overexpression of the target of selective pressure as well as detoxification, increased turnover or reduced uptake of the selective compound. While a selection is likely to generate mutations changed in a specific response or event, we have currently no means of predetermining the step in a process in which a mutation might be generated. We have used T - D N A tagging as the initial step in a genetic analysis of the molecular basis of the action of the plant growth substances auxin and cytokinin, as well as polyamines, substances that have long been implicated as playing an important role in plant growth and development. Such an approach is likely to be fruitful in these areas of research because, despite intensive biochemical analysis, little is known of the molecular basis of auxin/cytokinin action and the exact role of polyamines in plant growth and development remains a matter of debate. Here, we have summarized our initial results. Currently, we have studied in some detail about 10 mutants selected for growth in the absence of auxin, five selected for growth in the absence of cytokinin and eight selected for growth in the presence of M G B G . To ease analysis, we have only studied in detail those that contain single- or low-copy T - D N A inserts. Although still at a very preliminary stage, several points emerged. 1. Flexibility of the tagging event. We deliberately used transcriptional enhancers lacking a transcriptional start site so that in the activation of expression of plant genes we were released from the need for the T - D N A to insert upstream from an open reading frame which had no intervening start codons. This precluded us from being able to (1) generate antisense mutations following insertion of the tag downstream from a target gene, (2)recover the tagged gene from the plant genome by inverse PCR. In practice, however, it has allowed us to recover overexpression mutants where the tag has inserted upstream or downstream form the target gene. For example, in the case o f a x i 159 we know that the tag has inserted downstream from axi I ca. 6 kb from the transcriptional start site. In addition, preliminary analysis of several of the other tagged mutants that have been obtained indicates that the region [290]

of the plant genome responsible for the selected phenotype does not necessarily need to flank the right-border sequence of the T-DNA. 2. Cells able to grow in culture in the absence of auxin or cytokinin are able to regenerate into plants. The selection for growth in the absence of growth substances is based firmly on the observations, dating back at least 50 years, that neoplastic growth as a result of Agrobacterium transformation, or habituation, occurred under similar conditions. Such cultures are unable to regenerate into plants and at the outset we thought that the same would be true in our experiments. Although we have generated cell lines that have difficulty in regenerating, the majority have regenerated into plants able to progress through a normal growth cycle. While it may be premature to reach definitive conclusions before the functions of the tagged gene products are identified, it does suggest that the plant cell may have a variety of default systems so that overexpression of these genes does not disrupt normal developmental pathways. Thus, the types of gene tagged in these experiments contrast with the master genes of specific developmental pathways, most notably these of flower development, which apparently tolerate no such default pathways. 3. Dominant mutations ease functional testing. The ultimate proof that a specific gene has indeed been tagged and rescued is functional rescue. With recessive mutations this involves complementation of the mutant line with the wild-type allele of the mutated gene. In the case of dominant mutations, as described here, this involves the transformation of wild-type protoplasts with rescued plasmid sequences followed by screening for growth under selective conditions. This has proved to be a rapid and simple means of not only judging whether a gene has been rescued but also determining with deletion derivates where the gene is located. Routinely, a result is obtained 2 to 3 weeks at the most after protoplast transformation. In this review we have described our preliminary work involving activation tagging. We think that this form of T - D N A tagging may provide a means of isolating genes involved in complex bio-

1527 chemical and morphological pathways. The selection schemes that we have adopted were deliberately aimed at yielding mutants affected in processes that are currently little understood at the molecular level. While in the short term our goal is to characterize the action of the genes that we have isolated, the numbers of apparently differing mutants that we have generated should allow us, in the long term, to dissect genetically the systems underlying the action of plant growth substances and polyamines.

Acknowledgements We thank the undergraduate students who worked with the principle author in establishing the conditions for the experiments described here: Ursula Uwer, Michaela Dehio and Christel Schipmann. We have been technically assisted by Inge Czaja and Elke Bongartz. The work described had been funded by the Max Planck Gesellschaft, the Alexander von Humboldt Stiftung and the Graduiertenkolleg of the University of Cologne.

References 1. Cline MG: The role of hormones in apical dominance. New approaches to an old problem in plant development. Physiol Plant 90:230-237 (1994). 2. Davies PJ (ed): Plant Hormones and their Role in Plant Growth and Development. Martinus Nijhoff, Dordrecht (1987). 3. Depicker AG, Herman L, Jacobs A, Schell J, Van Montagu M: Frequencies of simultaneous transformation with different T-DNA and their relevance to Agrobacterium/ plant cell interaction. Mol Gen Genet 210:477-484 (1987). 4. Estelle M: The plant hormone auxin: Insight in sight. Bioessays 14:439-444 (1992). 5. Evans PT, Malmberg RL: Do polyamines have roles in plant development? Annu Rev Plant Physiol Plant Mol Biol 40: 235-269. 6. Feldman K: T-DNA insertional mutagenesis inArabidopsis: mutational spectrum. Plant J 1:71-82 (1992). 7. Gibson, S, Somerville CR: Isolating plant genes. Trends Biotechnol 11:306-313 (1993). 8. Goldsmith MHM: Cellular signalling: new insights into the action of the plant growth hormone auxin. Proc Natl Acad Sci USA 90:11442-11445 (1993).

9. Hayashi H, Czaja I, Schell J, Walden R: Activation of a plant gene implicated in auxin signal transduction by T-DNA tagging. Science 258:1350-1353 (1992). 10. Jacobs WP: Plant Hormones and Development. Cambridge University Press, Cambridge (1979). 11. Jones AM: Surprising signals in plant cells. Science 263: 183-184 (1994). 12. Kay R, Chan A, Daly M, McPherson J: Duplication of CaMV 35 S promoter sequence creates a strong enhancer for plant genes, Science 236:1299-1302 (1987). 13. King PJ: Plant hormone mutants. Trends Genet 229: 181-188 (1988). 14. Klee H, Estelle M: Molecular genetic approaches to plant hormone biology. Annu Rev Plant Physiol 42:529-551 (1991). 15. Koncz C, Nrmeth K, Rrdei G, Schell, J: T-DNA mutagenesis in Arabidopsis. Plant Mol Biol 20:963-976 (1992). 16. Koncz C, Martini N, Mayerhofer R, Koncz-Kalman Z, KOrber H, Rrdei GP, Schell J: High-frequency T-DNA mediated gene tagging in plants. Proc Natl Acad Sci USA 86:8467-8471 (1989). 17. Koncz C, Schell J: The promoter of the TL-DNA gene 5 controls the tissue specific expression of chimeric genes carried by a novel type of Agrobacterium binary vector. Mol Gen Genet 204:383-396 (1985). 18. Linsmaier EL, Skoog F: Organic growth factor requirements of tobacco tissue cultures. Physiol Plant 18: 100127 (1965). 19. Malmberg RL, Mclndoo J: Ultraviolet mutagenesis and genetic analysis of resistance to methlyglyoxal-bis(guanylhydrazone). Mol Gen Genet 196:28-34 (1984). 20. Malmberg RL, Mclndoo J, Hiatt AC, Lowe Ba: Genetics of polyamine biosynthesis in tobacco: developmental switches in the flower. Cold Spring Harbor Symp 50: 475-482 (1985). 21. Malmberg RL, Rose DG: Biochemical genetics of resistance to M G B G in tobacco: mutants that alter SAMdecarboxylase or polyamine ratios, and floral morphology. Mol Gen Genet 207:9-14 (1987). 22. Marsh JL, Erfle M, Wykes, EJ: The plC plasmid and phage vectors with versatile cloning sites for recombinant selection by insertional inactivation. Gene 32:481-485 (1984). 23. Marton L, Wullems GJ, Molendijk L: In vitro transformation of cultured cells from Nicotiana tabacum by Agrobacterium tumefaciens. Nature 277:129-131 (1979). 24. Murashige T, Skoog F: A revised medium for rapid growth and bioassays with tobacco tissue cultures. Physiol Plant 15:473-497 (1962). 25. Nagata T, Takebe I: Cell wall regeration and cell division in isolated tobacco mesophyll protoplasts. Planta 92: 301-308 (1990). 26. Napier RM, Venis MA: From auxin binding protein to plant hormone receptor? Trends Biol Sci 1 6 : 7 2 - 7 5 (1991).

[291]

1528 27. Odell JT, Nagy F, Chua N-H: Identification of DNA sequences required for the activity of the cauliflower mosaic virus 35S promoter. Nature 313:810-812 (1985). 28. Spena A, Estruch JJ, Schell J: On microbes and plants: new insights in phytohormonal research. Curt Opin Biotechnol 3:159-163 (1992).

[2921

29. Theologis A: Rapid gene regulation by auxin. Annu Rev Plant Physiol Plant Mol Biol 37:407-438 (1986). 30. Walden R, Hayashi H, Schell J: T-DNA as a gene tag. Plant J 1:281-288 (1991). 31. Walden R, Fritze K, Harling H: Induction of signal transduction pathways through promoter activation. Meth Cell Biol Plant Cell Biol, in press (1994).