PlantCell Reports

Plant Cell Reports (1989) 7:611-614

© Springer-Verlag 1989

Recovery of transgenic plants from "escape" shoots Alan McHughen and Mark C. Jordan Crop Development Centre, University of Saskatchewan, Saskatoon, Sask., S7N 0W0, Canada Received October 13, 1988/Revised version received November 21, 1988 - Communicated by A. R. Gould

Abstract

The problem of escapes is well known to those investigating the regeneration of transgenic shoots from transformed callus. Shoots can pass various tests and assays for transformation, and are then scored as transgenic, but the progeny do not express the transferred trait and do not contain the T-DNA. Explanations for these enigmatic "escapes" include instability of the T-DNA, genomic rearrangements during meiosis, or merely non-rigorous selection or identification assays giving rise to spuriously positive scorings. At least some shoots, however, are likely to simply be chimeric, containing both transformed and non-transformed cell lines. In this case, the transformed cells are responsible for the positive selection and scoring on tests, but either do not contribute to the germ line (resulting in no transgenic progeny) or contribute to only a portion of the germ line (resulting in many fewer positive segregants than expected). We describe two methods which we used to recover fully transgenic plants from apparent escapes. One method involved analyzing more progeny than would normally be necessary (to identify minority transgenic contribution to the cell line). The other method, (to recover transgenic plants from primary selectants with no transgenic contribution to the germ line) involved regenerating new shoots from leaf tissue used in a selection assay to score the initial shoot as a positive transgenic.

Introduction

The production of transformed callus is relatively easy in most dicotyledonous species attempted, through the use of either Agrobacterium vectors, electroporation, or other methods of DNA delivery into cells. The subsequent recovery of transgenic shoots (and, ultimately, whole plants) from transformed callus is a different matter; there a r e still only a few examples of species from which whole transgenic plants have been produced (including some major crops such as soybean (Hinchee et al., 1988; MeCabe et al., 1988), canola (Pua et al., 1987; Fry et al., 1987; Charest et al., 1988), flax (Jordan and McHughen, 1988a), and tomato (McCormick et al., 1986; Fillatti et al., 1987)).

Offprint requests to: A. McHughen

In most systems where transgenic plants can be recovered there is a common problem with the incidence of escapes: shoots that regenerate in selection medium from clearly transformed callus, often passing such tests for transformation as nopaline assays or even Southern analysis (Jordan and McHughen, 1988b; Horsch et al., 1988) but turn out negative on the definitive test, that of progeny segregation. The usual explanation for escapes is that the shoots arise from non-transformed cells in the callus mass (which is a mixture of transformed and non-transformed cells), that nopaline is transported into the shoots from nearby transformed cells (accounting for the positive nopaline response (Jordan and McHughen, 1988b)), that the surrounding transformed cells protect the non-transformed shoot initials from the effects of the selection agent by detoxifying it before it gets to the sensitive cells (in the case of NPT II), that contaminating bacteria (in the case of Agrobacterial transformation studies) provide the source for positive assay results, or that T-DNA is specifically rearranged or deleted between the time of the assay in the regenerated shoot and the assay on the progeny. In our experiments with flax (Linum usitatissimnm L.) transformed with any of several disarmed Agrobacterium Ti strains, we found that at least some of the shoots which would typically be labeled escapes are likely to be chimeric, containing a portion of transformed and a portion of non-transformed cells. By using a leaf callus on selection medium as the source tissue for regeneration, or by analyzing more progeny than should (theoretically) be necessary, we have recovered transgenic plants with progeny that express the transferred trait. Materials

and

Methods

Flax (Linum usitatissimum L. cv. "McGregor") seeds germinated aseptically and the hypocotyl excised and transferred to a petri dish containing Murashige and Skoog (1962) basal medium plus 1 mg/l benzyladenine and 0.02 mg/l naphthalene acetic acid, solidified with 0.8% agar. Hypocotyls were maintained for 10 days on this medium prior to inoculation with the bacteria. At inoculation, the hypocotyls were sliced to provide were

612 injury necessary for bacterial attachment and activation. Inoculated hypocotyls were returned to the petri dishes with medium of the same composition as above for two days, then transferred to similar medium but containing 400 mg/1 kanamycin, 250 mg/l Cefotaxime and 500 mg/l carbenicillin. Cultures were maintained on a light bench with a 16:8 hr day:night regime; temperature was constant at about 20°C. Callus formed on most of the inoculated hypocotyls and shoots arose from most of the calli. When shoots had several leaves they were excised from the callus and placed in vials containing MS basal medium for further development and root formation. At the time of excision, at least two leaves were removed from the shoot and placed on MS basal medium with 1.0 mg/l benzyladenine, 0.02 m g / 1 N A A , and i00 mg/l kanamycin. If the leaves callused, the callus was transferred to regeneration medium (MS basal plus 1 mg/l BA and 0.02 mg/l NAA), the originating shoot was scored positive, micropropagated and transferred to the greenhouse, in soil, to mature. Seeds were ultimately harvested and germinated on MS basal medium plus 500 mg/l kanamycin. Any shoot regenerated from the leaf callus was treated the same way. For micropropagation, established shoots in vials (about 3 cm tall) were excised halfway down, with the shoot tip transferred to a fresh vial to root and establish; the initial shoot typically regenerated axillary shoots to recover. This procedure was repeated several times to increase the number of plants from a single transformation event. The bacteria used in this work was Agrobacterium tumefaciens A208 carrying the disarmed pTiT37ASE (Rogers et al., 1987) cointegrated with the pMON570 intermediate construct which contains a chimeric NPT-II gene, a wild-type nopaline synthase and a 5-enolpyruvylshikimate3-phosphate synthase gene originally isolated from a petunia overproducer cell line (Shah et al., 1986) and further modified to confer enhanced glyphosate tolerance. These constructs were provided courtesy of Monsanto Co., St. Louis, USA. The other strain was C58 containing pGH6 (described by Haughn et al., 1988), which carries a mutant Arabidopsis ALS gene resistant to the herbicide ehlorsulfuron. This construct was provided courtesy of George Haughn. After the regenerant plants flowered and set selfed seed, the seeds were harvested and germinated on MS basal medium with 500 mg/1 kanamycin or InM chlorsulfuron, as appropriate, in Magenta pots. All viable seeds germinate at these levels, but wild-type segregants and controls produce bleached true leaves on 500 mg/l kanamycin, and stunted root development in the i nM chlorsulfuron. Transgenie segregants show normal green leaves and normal root growth, respectively.

Results

Fig. I. Leaf piece callusing on i00 mg/l kanamycin, with regenerant shoot (arrow). The original shoot was scored as a putative transgenic, while shoots providing other leaves in this plate were all negative.

None of the seed progeny of the putative transgenic shoot expressed the transgenic traits, so the original shoot was regarded as an escape. raised to maturity and set selfed seed. None of nine tested plants produced any progeny resistant to kanamycin or glyphosate according to the germination test (Fig. 2; Table i). Ordinarily, the original plant would be then scored as an escape and disregarded, but because a shoot was produced from the leaf callus on kanamycin we decided to maintain and further test that line. A total of three shoots were regenerated from the leaf-callus, and these were also micropropagated. The progeny of two of the shoots from the leaf callus segregated for kanamycin and/or glyphosate resistance in germination tests, while the third was negative (Table i). This result clearly shows that at least part of the initial regenerant was composed of transformed cells because the leaf which callused on kanamycin must have been at least partially composed of transformed cells; if all of the transformed cells remained in the callus, they would not be able to protect the leaf tissue from the kanamycin in this test. The proportion of transformed cells in the initial shoot was probably low, because there was apparently none in the germ'line, even after considerable subdivision through micropropagation. The proportion of transformed cells in the leaf callus must have been much higher, because two of three regenerants from that callus were transgenic and did pass the introduced genes to the progeny in a normal Mendelian manner. The organized structure of the leaf might provide a more even exposure of the cells to the selection agent, thus reducing the degree of cross protection so evident in the initial hypocotyl callus selection system (Jordan and McHughen, 1988b).

and Discussion

Two leaf pieces were taken from a shoot arising from an inoculated hypocotyl and transferred to the kanamycin selection medium. One piece callused (and spontaneously produced a new shoot) while the other did not (Fig. i). The original shoot was scored as a putative transgenic and micropropagated; a total of 16 cloned plants were

Fig. 2. Leaf pieces from positive chlorsulfuron-resistant segregant (left and right) and from negative (chlorsulfuron-sensitive) segregant (center) on 400 nM chlorsulfuron.

613 Table I. Progeny ratios for seeds germinated on 500 mg/l kanamycin (K) or lmM glyphosate (G), from plants from original shoot (#i-9) and from the three shoots (A,B,C) regenerated from the leaf callus on kanamycin.

source plant # select, medium pos. neg.

original shoot*

leaf callus**

1

2

3

4

5

K 0 20

K 0 9

G 0 14

G 0 15

G 0 15

6

7

8

9

AI

A2

A3

A4

A4

A5

G G G 0 0 0 19 17 20

G 0 20

K 7 2

G 16 5

G 13 6

K 7 4

G Ii 5

K 4 4

A5

B

C1

C1

G K 11 0 7 19

K 8 3

G 12 7

*-plants # 2-9 were micropropagated from plant i. **-leaf callus shoots are designated A,B,C; A2-A5 were micropropagated

Segregation ratios from some other transgenic regenerants support the interpretation that regenerating shoots can be chimeric. When transferring and assaying for single dominant traits in transgenic plants, one expects to find normal Mendelian ratios (3:1, 1:2:1, etc., depending on the type of expression) so long as the insert is into the nuclear genome. Often, however, non-Medelian ratios are encountered, with both higher than expected transfer (indicating multiple insertions or sometimes cytoplasmic genome insertions) and lower than expected transfer. Elaborate scenarios have been forwarded to explain the latter, such as complex genomic rearrangements during meiosis, or general instability of the transferred construct, as well as inefficient selection or assay protocols. While these reasons will account for many escapes, the simplest explanation is that the transformed cell line contributed to only a portion of the gametes, as would be expected from a chimeric regenerant. In another experiment, a flax plant was regenerated from hypocotyl callus formed after inoculation with a vector carrying a modified Arabidopsis ALS construct resistant to the herbicide chlorsulfuron (Haughn et al., 1986). The plant was taken directly through to maturity (not recycled through leaf callus on selection medium) and set selfed seed. The progeny segregated 1:15 (n = 2 resistant, 28 sensitive), according to the seed germination on inM chlorsulfuron test. In the first germination test, I0 seeds from the regenerant were used; none was significantly less inhibited than the others. In the second test, 20 seeds were germinated on InM chlorsulfuron. Two of these were better than the others (n = 18, mean root length of 1.37 cm + 0.15; no. 24 root length = 6.50 cm, no. 30 root length = 3.25 cm). Confirmation of chlorsulfuron resistance in these two plants was established by leaf callus in 400 nM chlorsulfuron assay (Fig. 2). The progeny were also tested for nopaline (NOS was also present in the T-DNA). So, of the 30 progeny tested, 2 were resistant to chlorsulfuron and produced nopaline; 28 were sensitive to chlorsulfuron and did not produce nopaline. This result indicates that the transformed cell line in the original regenerant contributed to about one-forty-fifth of the gametes (McHughen et al., in prep.). While it is not valid to extrapolate to suggest that less than 2% of the regenerant was composed of the transformed cell line, it is fair to say that the shoot was chimeric, composed primarily of non-transformed cells. With the transferred gene here being a selectable dominant marker, as few as

from AI

8-9 seed progeny need be tested to achieve a 99% confidence in identifying at least one positive segregant (assuming normal 3:1 Mendelian segregation in a non-chimeric regenerant). If that number were tested in this case, the chances are that no positives would turn up and the regenerant scored as an escape. The leaf callus on selection medium assay seemed to be quite dependable, in that shoots providing leaves which callused also produced seeds which segregated for the trait (Jordan and McHughen, 1988a), but some apparent escapes were also noted. Occasionally, we observed callusing on one leaf, or a portion of one leaf, but not on the other (Fig. 3). This is not particularly surprising, because, although it is often assumed that shoots develop from a single cell and, therefore, all cells in the shoot should have the same genotype, this may not be a valid assumption. Shoot organogenesis from callus can involve many initial cells, unlike regeneration by somatic embryogenesis, which does appear to involve only one initial cell. To test our hypothesis that at least some regenerating shoots are of multicellular origin, we are currently transforming flax using a GUS construct and will test the regenerating shoots histologically for GUS activity.

Fig. 3 . A leaf callus assay on inM glyphosate where one portion of the leaf dies and another portion grows, indicating the likelihood of a chimeric regenerant°

614 Conclusions Our results have shown that "escape" regenerants are not always due to incorrect selection or identification assays, but that at least some apparent escape shoots have stable transformed cells in them. We can recover true transgenic plants from these cells by a second cycle of regeneration on selection medium, using leaf tissue from the original regenerant shoot as source tissue. A second method of recovery of transgenic plants is to test larger numbers of progeny than is usually necessary for identifying mendelian ratios. This method, however, requires at least some contribution by transformed cells to the germ line. The chimeric condition of the shoots could be due to inactivation of T-DNA in some of the cells early in shoot differentiation, thus leading to non-expression of transferred genes in all cells derived from them. This inactivation could be due to genetic rearrangements or deletions in the T-DNA. Alternatively, the chimeras could be due to a multicellular origin of the shoot, incorporating both transformed and non-transformed cells. In either case, the methods presented here offer an opportunity to salvage transgenic plants from at least some escapes. We believe that this phenomenon is due to a multicellular origin of the initial regenerant shoot, incorporating both transformed and non-transformed cells, in at least some cases. This hypothesis is currently being tested using histological analyses o~ GUS-transformed flax hypocotyl callus regenerant shoots.

Reference~

Haughn G.W., Smith J., Mazur B., Somerville C. (1988) Mol. Gen. Genet. 211: 266-271. Hinchee M.A.W., Conner-Ward D.V., Newell C.A., McDonnell R.E., Sato S.J., Gasser C.S., Fischhoff D.A., Re D.B., Fraley R.T., Horsch R.B. (1988) Bio/Technology 6: 915-922. Horsch R.B., Fry J., Hinchee M., Delannay X., Shah D., Kishore G., Fishhoff D., Tumer N., Klee H., Beachy R., Rogers S., Fraley R. (1988) XVI Int'l. Cong. Genet., Toronto, Canada. Jordan M.C., McHughen A. (1988a) Plant Cell Reports 7: 281-284. Jordan M.C., McHughen A. (1988b) Plant Cell Reports 7: 285-287. McCabe D.E., Swain W.F., Martinelli B.J., Christou P. (1988) Bio/Technology 6: 923-929. McCormick S., Niedermeyer J., Fry J., Barnason A., Horsch R.B., Fraley R.T. (1986) Plant Cell Reports 5: 81-84. Murashige T . , S k o o g F. (1962) Physiol. Plant. 15: 473-497. Pua E.C., Mehra-Palta A., Nagy F., Chua N.H. (1987) Bio/Technology 5: 815-817. Rogers S., Klee H.J., Horsch R.B., Fraley R.T. (1987) Methods Enzymol. 153: 253-277. Shah D.M., Horsch R.B., Klee H.J., Kishore G.M., Winter J.A., Tumer N.E. Hironaka C.M., Sanders P.R., Gasser C.S., Aykent S., Siegel N.R., Rogers S.R., Fraley R.T. (1986) Science 233: 478-481.

Charest P.J., Holbrook L.A., Gabard J., lyer ~.N., Miki B.L. (1988) Theor. Appl. Genet. 75: 438-445.

Acknowledgements Fillatti J.J., Kiser J., Rose R., Comai L. (1987) Bio/Technology 5: 726-730. Fry J., Barnason A., Horsch R.B. (1987) Plant Cell Reports 6: 321-325.

This project was supported by the Saskatchewan Agriculture Development fund and the Western Grains Research Foundation. We also thank Ms. Gina Feist for technical expertise.

Recovery of transgenic plants from "escape" shoots.

The problem of escapes is well known to those investigating the regeneration of transgenic shoots from transformed callus. Shoots can pass various tes...
575KB Sizes 0 Downloads 0 Views