Arabidopsis ecotypes and mutants that are ...

2 downloads 0 Views 3MB Size Report
the Agrobacterium attB mutant, and Dr Susan Karcher for critical reading of the ... Escudero, J., Neuhaus, G. and Hohn, B. (1995) Intracellular. Agrobacterium ...
The Plant Journal

(2000) 21(1). 9-16

Arabidopsis ecotypes and mutants that are recalcitrant to Agrobacterium root transformation are susceptible to germ-line transformation Kirankumar S. Mysore^^, C. T. Ranjith Kumar^ and Stanton B. ^Departmerit of Biological Sciences and ^Purdue Genetics Program, Purdue University, West Lafayette. IN 47907-1392, USA FiBceived 1 November 1999; accepted 9 November 1999. •For correspondence (fax +765 496 1496; e-mail [email protected]). 'Present address: Boyce Thompson Insltlute. Tower Rd, Cornell University, Ithaca, NY 14853, USA.

Summary Germ-line transformation (vacuum infiltration) is frequently used to transform Arabidopsis thaliana using Agrobacterium tumefaciens. We have recently identified several Arabidopsis ecotypes and T-DNA-tagged mutants that are recalcitrant to ^grobactenum-mediated transformation of cut root segments. Some of these ecotypes and mutants are deficient in their ability to bind bacteria. Some are deficient in T-DNA integration. We report here that using a germ-line transformation protocol we transformed these ecotypes and mutants, including attachment- and integration-defective Arabidopsis plants, with a frequency similar to that of highly susceptible wild-type plants. However, we could not transform otherwise highly susceptible Arabidopsis plants by germ-line or root transformation using several vir and attachment-deficient Agrobacterium mutants. These results indicate that certain plant factors important for transformation may exist in germ-line tissue but may be lacking in some somatic cells.

Introduction

Agrobacterium tumefaciens processes a portion of DNA, the T-(transferred) DNA, from the resident Ti- (tumor inducing} plasmid and transfers it to a plant cell where it may become stably integrated into the plant genome (Chilton etal.. 1977). During the past two decades scientists have studied intensively the molecular events that occur within the bacterium during this process. These investigations have resulted in the identification of bacterial virulence {vir) genes and an elucidation of the roles these genes play in T-DNA processing and transfer (for recent reviews see Sheng and Citovsky, 1996; Zupan and Zambryski, 1995; Zupan and Zambryski, 1997). In contrast, not much is known about plant-encoded factors involved in T-DNA transfer, nuclear targeting and integration. Scientists have enumerated several steps that must take place in plants during the transformation process (Sheng and Citovsky, 1996; Zupan and Zambryski, 1997). These steps include bacterial attachment to the plant cell, T-DNA transfer across the plant cellular membrane, nuclear translocation of the single-stranded T-DNA molecule (the 'T-strand'), and T-DNA integration into the plant genome. There is, however, little information regarding factors that may limit the stable transformation of particular cells ©2000 Blackwell Science Ltd

within a plant, or why certain species (or cultivars/ecotypes within a species) may differ in their susceptibility to transformation. Significant variation may exist in the transformation efficiency of different cultivars/ecotypes of a species (Nam etal., 1997), or tissues or cell types within a plant. As a result of this variation, scientists have utilized different tissues to obtain high levels of transformation. However, the choice of tissue differs among various species of plants (Van Wordragen and Dons, 1992). We currently lack an understanding of the source(s) of this variation. The identification of plant factors or genes that are involved in the T-DNA transformation process may help further our knowledge in this field and develop strategies for transforming recalcitrant plant species. Agrobacterium germ-line transformation (also known as vacuum infiltration or in planta transformation) has become one of the most popular methods for genetic transformation of Arabidopsis thaliana. The major reasons for the increased use of this method are its simplicity and the lack of requirement for tissue culture regimes. Transformants are often obtained at high frequency and in less time compared with other methods. Consequently, this method may not only be used to introduce transgenes.

10 Kirankumar S. Mysore et al. but may also effect random mutagenesis or T-DNA tagging (Azpiroz-Leehan and Feldmann, 1997}. Germ-line transformation, as commonly practiced at present, is a modification of an earlier vacuum infiltration method described by Bechtold etal. (1993). These authors infiltrated Arabidopsis plants with Agrobacterium and replanted them individually for the subsequent collection of seeds. Using current germ-line transformation protocols, Agrobacterium is simultaneously applied directly to numerous mature flowering plants, and seeds are collected after a few weeks (Bent and Clough, 1998).

germ-line transformation process may not need some of the plant factors required during conventional root inoculation. In the following study we show that Agrobacterium mutants that are avirulent by 'conventional' infection methods are also avirulent when used for germ-line transformation. However, we could readily transform, by germ-line transformation, many-4rab/dops/s ecotypes and T-DNA insertion mutants that were highly recalcitrant to transformation using conventional root transformation protocols.

Although germ-line transformation is a commonly used method to transform Arabidopsis, the origin of the resulting transformants remains obscure. The majority of the transgenic progeny obtained in the initial selection after germ-line transformation are hemizygous rather than homozygous for the transgene (Bechtold etal., 1993; Chang etal., 1994). This result suggests that transformation usually occurs late in floral development, after the formation of male and female gametophytes. Recently three laboratories reported that female reproductive tissue (the egg cell) is the target of T-DNA integration during in plar)ta transformation (Ye etai. 1999; abstract 279 (Bechtold etal.) and abstract 285 (Desfeux etal.), unpublished data, from the 9th International Conference on Arabidopsis Research, Madison, Wl, in 1998),

Results

We have identified several ecotypes and T-DNA-tagged mutants of Arabidopsis that are deficient in Agrobacteriummediated transformation of cut root tissue (Nam etal., 1997; Nam etal., 1999). In the process of attempting to complement the T-DNA-tagged mutants, we discovered that some of the mutants that are recalcitrant to transformation by inoculating cut root or flower bolt tissue can be transformed easily using a germ-line transformation procedure. This result suggested to usthatthe/^grodacter/um

Effect of different Agrobacterium virulence genes on germ-line transformation To determine the importance of various vir genes for germ-line transformation, we used several mutant Agrobacterium strains containing the T-DNA binary vector pBISNI (Narasimhulu etal., 1996) and harboring Tn3HoHol insertions (Stachel and Nester, 1986) in either virBl (mx243; At805) or virE2 (mx341; At807). pBISNl and the derivative plasmid pBlSN2 contain a nos-npf//chimeric gene and a gusAintron gene under the transcriptional regulation of a strong 'super-promoter' (Ni etal., 1995), We also utilized a virD2 mutant bacterial strain, containing the T-DNA binary vector pBISN2 (Narasimhulu etat.. 1996), that contains a w deletion/substitution mutation (At829; Mysore etal., 1998). Similar mutant Agrobacterium strains are either avirulent or extremely attenuated in virulence (Mysore etal., 1998; Shurvinton et at., 1992; Stachel and Nester, 1986). In addition to these bacterial strains, we introduced pBISNI into several attachment-deficient Agrobacterium mutants including attB' (Matthysse etat., 1996; C58::N004; At1064), chvA' (Douglas etal., 1985; At106a) and c/ivS" (Douglas etal., 1985; At1069). Finally,

Table!. Comparison between conventional root transformation and germ-line transformation of Arabidopsis using different Agrobacterium mutants Agrobacterium strain

Ti pissmid

% kan' calli from Amount (mg) of seeds germinated root infection from vacuum-infiltrated plants" (n)

At789 At790 At793 At805 At807 At829 At1063 At1064 At 1068 At1069

pTiA6 89 ±6 pTiBo542 92 ± 2 No pTi 0 pTiA6 virBT 0 pTiA6 virE2' 0.75 ± 2 pTtA6 virD2~ (Aw) 9 ±2'' pTiA6 + osa 2±3 pTiC58 lattBi 0 pTiAB ichvA) 0 pTiA6 ichvB) 3±2

780 (3) 790 (2) 660 (2) 590 (2) 610 (2) 460 (2) 610 (2) 910 (2) 2070 (3) 2180 (3)

Number of kanamycin- % transformed resistant plants seeds 55 39 0 0 0 0 0 0 0 0

0.15 ±0.02 0.11+0.02 0 0 0 0 0 0 0 0

^10 mg equals approximately 500 seeds. All seed lots showed greater than 95% germination. "Small calli. n, number of times the experiment was repeated with different batches of plants and bacteria. % kan^ percentage root segments yielding kan*^ calli. ©Blackwell Science Ltd, The Plant Jourr^al, (20001, 21, 9-16

Germ-line transformation of Arabidopsis we introduced pBISNl into an Agrobacterium strain containing the oncogenic suppresser gene osa (Close and Kado, 1991; Farrand etal.. 1981; Lee etal.. 1999; At10631. All these Agrobacterium mutants and wild-type strains (Table 1) have the same chromosomal background (C58). We used these Agrobacterium strains, as well as wildtype strains, to transform Arabidopsis plants (ecotype Ws) using a conventional in vitro root inoculation procedure INam etal.. 1997). We inoculated small sterile root segments of individual 3-week-old Arabidopsis plants with these Agrobacterium strains and selected kanamycinresistant calli derived from these root segments, as described previously (Nam etal.. 1997). The results of

11

these inoculations are shown in Table 1 and Figure 1(a). Using A. tumefaciens At789 and At790 (which contain wild-type wrand attachment genes), approximately 90% of the inoculated root bundles were transformed to produce kanamycin-resistant calli. In contrast, all A. tumefaciens virulence and attachment-deficient mutants tested, as vwell as A. tumefaciens At1063 (containing the osa gene) and At793 (lacking a Ti-plasmid), were either avirulent or extremely attenuated in virulence using the root transformation method. We attempted germ-line transformation of Arabidopsis with these same Agrobacterium strains. We germinated Arabidopsis seeds (ecotype Ws) in soil and grew them at 25''C for 3-4 weeks until they started to flower. Some fastgrowing flower bolts were clipped to maintain uniform flowering. Once flowers were seen on almost all the flower bolts, we dipped these flower bolts into the appropriate Agrobacterium culture and applied a vacuum (see the Experimental procedures for details). Seeds were allowed to set and were subsequently germinated on B5 medium containing kanamycin. Transgenic plants resistant to kanamycin grew rapidly, whereas kanamycin-sensitive seedlings turned yellow. Kanamycin-resistant plants were removed and stained with X-gluc. Occasionally a few kanamycin-resistant plants did not stain blue. These plants could be 'escape' plants or a transgenic plant with rearranged T-DNA or an inactive gusA gene. Only the seedlings that developed blue color (i.e. expressed the gusA gene) were scored as shown in Table 1. These results were similar to those of the root transformation: other than the wild-type Agrobacterium strains At789 and At790, all mutant Agrobacterium strains tested were avirulent using the germ-line transformation assay. In addition, osa inhibited transformation using this assay. These experiments were repeated at least twice on separate sets of plants on different dates with similar results (Table 1). These results suggest that germ-line transformation requires virulence and attachment genes of Agrobacterium.

Germ-line transformation of different ecotypes and mutants of Arabidopsis that are recalcitrant to conventional root transformation Rgurei. Transformation of Arabidopsis root segments with Agrobacterium tumefaciens. la) Root segments of wild-type Arabidopsis (ecotype Wsl were infected with wild-type A. tumefaciens At789 and various mutant bacterial strains ichvA, At1068; chvB, At1069; attB. At1064; virE2. At807) containing pBISNI, After co-cultivation for 2 days, the root segments were washed and incubated on medium containing timentin and kanamycin as described in the Experimental procedures. The plates were photographed after 1 month. ibi Root segments of various Arabidopsis ecolypes were co-cuitivated with A. tumefaciens GV3101(pCASl). After 2 days, the root segments were washed and incubated on medium containing timentin and phosphinothricin as described in the Experimental procedures. The plates were photographed after 1 month.

©Blackwell Science Ltd, The Plant Journal, (2000), 21, 9-16

We have previously identified several ecotypes and T-DNA insertion mutants of Arabidopsis that are deficient in Agrobacterium-medi\ated transformation (Nam etai. 1997; Nam etal., 1999). The mutants rati and rat3, and the ecotypes Petergof and BI-1, are deficient in attaching agrobacteria (Nam etal.. 1997, 1999). The mutant raf5 and the ecotype UE-1 are most likely deficient in T-DNA integration (Mysore ef al.. 2000; Nam etai. 1997). The transformation block points for the mutant rat9 and the ecotypes Bla-2 and Dijon-G are not currently known. We

12

Kirankumar S. Mysore et al.

T8ble2. Comparison between conventional root transformation and germ-line transformation of different mutants and ecotypes of Arabidopsis Arabidopsis ecotype or mutant

Probable step blocked in T-DNA transformation

%ppt'" calli from root infection

Amount (mgt of seeds Number of germinated from ppt-resistant vacuum-infiltrated plants" (n) plants

% transformed seeds

Reference

Ws

None None Bacterial attachment Bacterial attachment Bacterial attachment Bacterial attachment T-DNA integration T-DNA integration Unknown Unknovwn Unknown

87 = 5 89=: 10 5±2 9i2 3±9 1 ±2 15r5

1000 {2) 1700 (2) 1270(2) 1490 (2) 390 (1) 260 (1) 920 (2) 1030 (2) 960 (2) 1550(2) 200 (1)

0.26 i:0.13 0.63 iO.38 0.22 ±0,06 0.22 ±0.11 0.14 0.16 0.22 ^0.07 0.34 ±0.05 0.21 ±0.11 0.30 ±0.16 0.21

Nam Nam Nam Nam Nam Nam Nam Nam Nam Nam Nam

Aa-0 rati rat3 Petergof BI-1 rat5 UE-1 rat9 Dijon-G Bta-2

12-6'' 6±4 10 ±10" 0±0

140 651

140 174

27 21

in 163 139 262 21

etai, etai. etai, etai. etai. etai. etai. etai. etai. etai. etai.

1999 1997 1999 1999 1997 1997 1999 1997 1999 1997 1997

*10 mg equals approximately 500 seeds. All seed lots showed greater than 95% germination. ''Extremely small calli. n, number of times the experiment was repeated with different batches of plants and bacteria. %ppt^; % root segments yielding ppt^ calli.

retested these mutants/ecotypes for stable transformation to form phosphinothricin(ppt)-resistant calli using an in vitro root inoculation procedure. We used phosphinothricin instead of kanamycin as a selectable marker because the TDNA-tagged mutants were already kanamycin resistant. We infected sterile root bundles from these plants with the non-tumorigenic strain A. fumefac/ens GV3101 containing the binary vector pCASI. pCASI contains a chimeric nosbargene and a gus>^gene under the transcriptional control of a modified super-promoter (Ni etai, 1995). Phosphinothricin-resistant calli were scored as shown in Table 2. The results of these assays correlated well with those of tumorigenesis assays (Nam etai, 1997; Nam etai, 1999). Almost 90% of the root bundles of the highly susceptible ecotypes Ws and Aa-0 were transformed to form pptresistant calli, whereas other ecotypes and mutants tested showed an extremely attenuated transformation phenotype (Figure lb). We next tested these ecotypes and mutants for germ-line transformation. Seeds from vacuum-infiltrated plants were germinated on B5 medium in the presence of phosphinothricin, and the resulting ppt-resistant plants were tested for expression of GUS activity by staining with X-gluc. All ecotypes and mutants tested were transformed by germlinetransformation with approximately the same frequency as wild-type highly susceptible plants (Table 2). For most of these ecotypes and mutants the experiment was repeated twice on a separate set of plants on a different date with similar results (Table 2). We repeatedly obtained similar results with the rati, rat3 and rat5 mutants using an Agrobacterium strain containing a different binary vector with hygromycin-resitance as the selectable marker (J. Nam and K.S. Mysore, unpublished data). These results suggest that germ-line tissue of Arabidopsis is more

susceptible to Agrobacterium tissue in all the Arabidopsis

transformation than root

mutants and ecotypes tested.

Discussion We have previously developed an in vitro root inoculation assay to test Arabidopsis thaliana for susceptibility or resistance to Agrobacterium transformation and, using this assay, identified several ecotypes and T-DNA-tagged mutants that are highly recalcitrant (Nam etai. 1997; Nam etai, 1999). In the course of attempting to complement several T-DNA-tagged mutants by transformation with the wild-type gene, we noticed that although they were highly recalcitrant to transformation by root inoculation, they were easily transformed by vacuum infiltration. We therefore conducted this study to determine, using vacuum infiltration, the relative transformation frequencies amongst Arabidopsis ecotypes and mutants that were recalcitrant to root transformation. We also investigated the frequency of transformation of highly susceptible ecotypes of Arabidopsis using both transformation methods and various Agrobacterium mutants. Our results indicated that the transformation frequency using either method with these mutant bacteria was negligible, Unexpectedly, all Arabidopsis mutants and ecotypes tested in this study that were recalcitrant to root transformation were highly susceptible to germ-line transformation. However, we have identified previously a mutant, rad5. that is recalcitrant to both root and germ-line transformation (Nam etai, 1998). This latter result indicates that the germ-line transformation protocol that we used will not succeed with all 'transformation-deficient' Arabidopsis lines. ©Blackwell Science Ltd, The Plant Journal

(2000), 2 1 , 9-16

Germ-line transformation of Arabidopsis 13 Until recently, root transformation was widely used as the best method to transform Arabidopsis (Grevelding etai, 1993;Valvekens etai. 1988). More recently, in plants! vacuum infiltration/germ-line transformation has become a popular method to transform Arabidopsis because of its simplicity and its capability to generate transformants at a high frequency (Bechtold etai, 1993; Bent and Clough, 1998), There are at least four steps in which plant factors are likely involved in the Agrobacterium-medialed transformation process: bacterial attachment to the plant cell surface; transfer of T-strands from the bacteria to plant cells across the plant cell wall and membrane; transport of the Tcomplex to the plant nucleus; and stable integration of TDNA into the plant genome. Our ability to transform by germ-line transformation some of the T-DNA-tagged Arabidopsis mutants that are deficient in Agrobacterium attachment (Nam etai, 1999) initially suggested to us that germ-line transformation may not require bacterial attachment. Lippincott and Lippincott (1969) first suggested that attachment of Agrobacterium to the plant cell surface is an essential step in tumorigenesis. All non-attaching Agrobacterium mutants are avirulent, as determined by standard tissue inoculation assays (Cangelosi etai, 1987; Douglas etai. 1982; Matthysse, 1987; Thomashow etai. 1987). To date, four genetic loci of Agrobacterium. chvA. chvB, pscA {exoQ and att, have been identified as playing a role in the attachment of Agrobacterium to plant cells (Broek and Vanderieyden, 1995). The identity of the bacterial receptor on the plant surface is not known. Earlier experiments showed that when the plant cell surface was treated with various proteinases before incubation with bacteria, bacterial attachment was inhibited, suggesting that a protein on the plant surface is required for this process (Gurlitz etai. 1987; Neff and Binns, 1985). Two proteins, a rhicadhesin-binding protein (Swart etai, 1994) and a vitronectin-like protein (Wagner and Matthysse, 1992), have been proposed as specific receptors for Agrobacterium attachment. Our identification of transformation-deficient mutants and ecotypes of Arabidopsis that are deficient in binding Agrobacterium (Nam etai. 1997; Nam etai, 1999) strongly suggests that plant factors are required for Agrobacterium-plant cell attachment. However, when Agrobacterium cells that have been induced previously with acetosyringone are microinjected into plant cells, they can transfer T-DNA to the nucleus even when the bacteria are attachment-deficient (Escudero etai, 1995). Therefore under these 'artificial' inoculation conditions, binding of bacteria to the plant cell and T-DNA transfer can become 'unlinked'. To test the effect of different Agrobacterium mutants on germ-line transformation we selected several mutants that are defective either in one of the virulence genes (virBI. ©Blackwell Science Ltd, The Plant Journal (2000), 21, 9-16

virE2 or the a region of virD2) or in genes {chvA . chvB~ or attB) that are involved in bacterial attachment to the plant cell. Our in vitro root inoculation experiment results showed that all these mutants are highly deficient in their ability to transform Arabidopsis plants to form kanamycinresistant calli. We did observe growth of a few antibioticresistant calli using some of these mutants, indicating that our assay is very sensitive (Table 1), Highly attenuated tumorigenesis resulting from infection by A tumefaciens strains harboring osa or mutations in virE2, the o) region of virD2. or chvB has been documented previously (Dombek and Ream, 1997; Farrand etai. 1981; Hawes and Pueppke, 1989; Shurvlnton ef ai, 1992). Our inability to transform Arabidopsis by germ-line transformation using attachment-deficient Agrobacterium mutants was unexpected and inconsistent with our initial hypothesis that germ-line transformation does not require bacterial attachment. However, the attachment-deficient plant mutants {rati, rat3) and ecotypes (Petergof, BI-1) of/^rafa/dops/s that were root transformation-deficient (Table 2) were easily transformable by germ-line transformation with almost the same frequency as that of the wild-type progenitor plants (ecotypes Ws and Aa-0). These latter results would tend to support our initial theory that bacterial attachment is not required for germ-line transformation. This apparent contradiction regarding the necessity of bacterial attachment to plant cells to effect germ-line transformation warrants further investigation. One possible explanation for the ability of these mutants and ecotypes of Arabidopsis, recalcitrant to Agrobacterium root transformation, to be transformed by germ-line transformation is the differential expression of plant factors in different tissues. We propose that root transformation (and perhaps transformation of all somatic tissues) requires 'factors' that are limiting in these tissues but that are abundant in the germ-line tissue. The abundance of these factors in the germ-line tissue may allow transformation of this tissue even in mutants or ecotypes that are recalcitrant to root transformation. These factors may be specific gene products, or may be active DNA metabolism that occurs during the process of meiosis. An example of how active DNA metabolism in the germ-line tissue may permit transformation of particular ecotypes or mutants may be found with the mutant rat5 and the ecotype UE-1. Both of these plant lines are recalcitrant to /Agrobacfer/um-mediated transformation of roots and flower bolts (Table 2; Nam etai, 1997; Nam etai. 1999; Mysore ef al, 2000). Transformation of UE-1 is blocked at the T-DNA integration step (Nam etai, 1997). We have recently shown that the rat5 mutant is also blocked at the T-DNA integration step: Agrobacterium-Tr\e