Plant Cell Reports
Plant Cell Reports (1996) 16:219-221
© Springer-Verlag1996
A novel principle for selection of transgenic plant cells: positive selection Morten Joersbo and Finn T. Okkels* Danisco Biotechnology, Langebrogade l, DK-1001 Copenhagen K, Denmark * Present address: Dansk Planteforaedling, Hojerupvej 31, DK-4660 Store Heddinge, Denmark Received 5 February 1996/Revised version received 8 May 1996 - Communicated by H. L6rz
Summary. A novel principle for selection of transgenic plant cells is presented. In contrast to traditional selection where the transgenic cells acquire the ability to survive on selective media while the non-transgenic cells are killed (negative selection), this selection method actively favours regeneration and growth of the transgenic cells while the non-transgenic cells are starved but not killed. Therefore, this selection strategy is termed 'positive selection'. The E. coli 13-glucuronidase gene was used as selectable (as well as screenable) gene and a glucuronide derivative of the cytokinin benzyladenine as selective agent which is inactive as cytokinin but, upon hydrolysis by GUS, active cytokinin is released stimulating the transformed cells to regenerate. Selection of Agrobacterium tumefaciens inoculated of tobacco leaf discs on benzyladenine N-3-glueuronide (7.5-15 mg/1) resulted in 1.7-2.9 fold higher transformation frequencies compared to kanamycin selection. A significant advantage of this selection procedure is the elimination of the need for herbicide and antibiotic resistance genes.
Abbreviations GUS - f3-glucuronidase, HPLC - High pressure liquid chromatography, NPT II - Neomycin phosphotransferase II, X-gluc - 5-bromo-4-chloro-3-indolylglucuronide.
Introduction For genetic transformation of plants, it is well known that only a minor fraction of the treated cells become transgenic while the majority of the cells remain untransforreed which is why they have to be eliminated by selection. The general strategy has been to introduce a selecCorrespondence to: M. Joersbo
table gene along with the gene(s) encoding the desired trait allowing the transformed cells to grow on selective media while the non-transformed cells are killed (Bowen 1993). The most widely used selectable gene is the neomycin phosphotransferase II (NPTII) gene (Fraley et al. 1986) which confers resistance to the aminoglycoside antibiotics kanamycin, neomycin and G-418 (Bevan et al. 1983). A number of other selective systems has been developed based on resistance to bleomycin (Hille et al. 1986), bromoxynil (Stalker et al. 1988), chloramphenicol (Fraley et al. 1983), 2,4-dichlorophenoxy-acetic acid (Streber and Willmitzer 1989), glyphosate (Shah et al. 1986), hygromycin (Waldron et al. 1985) or phosphinothricin (De Block et al. 1987). All of these compounds are antibiotics or herbicides. The corresponding resistance genes are in many cases not relevant to the desired transgenic trait and they may be undesirable after selection has been accomplished. Here we present a novel concept for selection of transgenic plant cells. The method exploits the fact that cytokinin must be added to plant explants in order to obtain optimal shoot regeneration rates. By adding cytokinin as an inactive glucuronide derivative, cells which have acquired the GUS gene by transformation are able to convert the cytokinin glucuronide to active eytokinin (Fig. 1) while untransformed cells are arrested in development. Because of the metabolic advantage the transgenic cells obtain, this is a positive selection strategy as opposed to traditional selection where the transgenic cells only become insensitive to the selective agent. In the positive selection system, the GUS gene serves the dual purpose of being both a selectable and screenable marker gene. Therefore, no herbicide or antibiotic resistance genes are needed.
220 Materials and methods Cytokinin glucuronide. Benzyladenine N-3-glucuronide was synthesized by Dr. R. Whenham, Apex Organics, Devon, UK (manuscript submitted) and was tested as a substrate for the E. coli GUS (Sigma G7896) by incubating a 1.0 mg/ml solution in 50 mM Naphosphate buffer pH 7.0 in the presence of GUS for 18 hours at 37°C. HPLC analysis (conditions modified according to Palmgren et al. 1990) showed a virtually complete removal of the benzyladenine N-3-glucuronide peak with concomitant production of a peak which cochromatographed with authentic benzyladenine. Transformation. Sterile shoot cultures of tobacco OVicotiana tabacum vat. 'Wisconsin 38') were maintained on MS-medium (Murashige and Skoog 1962). Leaf discs from the largest leaves (3-5 weeks old) were punched out and was transformed according to standard procedures using Agrobacterium tumefaciens strain LBA 4404 harbouring pBI121 containing both a GUS and NPTII gene (Jefferson et al. 1987). Selection. After co-culture for 3 days the leaf discs were transferred to selection medium (MS-medium) containing no cytokinin or kanamycin but various concentrations of benzyl-adenine N-3-glucuronide according to the experiment. In the controls, 1.0 mg/t benzyladenine and 300 mg/1 kanamycin sulfate were added instead ofbenzyladenine N-3-glucuronide. All selection media contained 800 rag/1 carbenicillin in order to eliminate the Agrobacteria. Shoot tips from regenerated shoots were tested for GUS activity by the X-gluc histological assay according to Jefferson et al. (1987). Only ieaf tips displaying strong blue reaction were evaluated as being positive. For each selection medium, 90-120 leaf discs were initiated in 3 or more independent experiments.
NH
CH2" ~
~----~l
N
Results
Incubation o f benzyladenine N-3-glucuronide with GUS resulted in total hydrolysis as evidenced by the release o f benzyladenine. This demonstrated that benzyladenine N3-glucuronide is substrate for GUS. Benzyladenine N-3-glucuronide was tested as a selective agent at various concentrations in the tobacco leaf disc transformation system and was found to give transformation frequencies which were superior to those obtained with kanamycin selection (Tab. 1). At 7.5 and 15 rag/1 benzyladenine N-3-glucuronide the transformation frequencies were 0.15 and 0.26 transformed shoot per explant, respectively, compared to 0.09 by kanamycin selection. Saccharo- 1,4-1actone is a potent inhibitor o f GUS (Fishman 1955). Addition of saccharo-l,4-1actone (at 1.92 g/l) reduced the transformation rate to 0.02 transformed shoots per explant indicating that the selection is dependent on the presence o f active GUS and that unhydrolysed benzyladenine N-3-glucuronide has only a very weak if any cytokinin activity. Saccharo-1,4-1actone up to at least 3.5 g/1 had no significant effect on the regeneration o f shoots from non-transgenic explants.
NH,, CH2
O
II
GUS
C~OH
lo ° H
HO
\
H
Glucuronic acid
( OH
Benzyladenine N-3-glucuronide
Benzyladenine
Fig. 1. Principle of positive selection. In cells transformed with a GUS gene the inactive cytokinin(benzyladenineN-3-glucuronide)is hydrolysedby GUS to active cytokinin (benzyladenine)which stimulates shoot regeneration and growth of the transgenic cells.
221 Benzyladenine N-3-glucuronide (mg/1)
Benzyladenine (mg/1)
Saccharo-l,4lactone (g/l)
Kanamycin sulfate (mg/1)
GUS+ shoots per leaf disc (+ SD)
7.5
0
0
0
0.15 (+ 0.11)
15.0
0
0
0
0.26 (_+ 0.14)
15.0
0
1.92
0
0.02 (_+ 0.01)
0
1.0
0
300
0.09 (_+ 0.04)
Tab. 1. Transformationfrequencies obtainedby positiveselection of tobacco shoots selected on media containing benzyladenine N-3-glucuronide in the presence or absence of saccharo-l,4-1actone, comparedto traditional selection on kanamycin (1.0 mg/1 benzyladenine and 300 mg/1 kanamycin sulfate). The transformation frequencies along with standard deviations (SD) were calculatedas the averagenumberof GUS-positiveshoots per leaf disc in 3 or more independent experiments. Discussion
We have presented a novel principle for selection oftransgenic plant cells, based on the fact that in vitro grown plant tissue cultures are auxotrophic in several aspects. In particular, exogenously applied cytokinin significantly enhances regeneration and growth. Due to the acquired ability to hydrolyse benzyladenine glucuronide, the transgenic cells harbouring the GUS gene were stimulated compared to the non-transgenic cells, hence the term 'positive selection'. Selection of transgenic plant cells by positive selection rather than negative selection offers significant advantages. Firstly, the transformation frequencies obtained by positive selection appear to be higher than those obtained with kanamycin selection. This could be related to the fact that during negative selection the majority of the cells in the explants die. Such dying cells may release toxic substances (such as phenolics) which in turn may impair regeneration of the transformed cells (Bowen 1993). In addition, dying cells may form a barrier between the medium and the transgenic cells preventing or slowing uptake of essential nutrients. The overall effect of necrotic tissue during selection is presumably reduced mitotic activity of the transformed cells resulting in less transgenic shoots emerging from the explants. When benzyladenine N-3-glucuronide was used as selective agent the explants maintained significantly higher viability during the selection procedure, compared to kanamycin selection. Secondly, no antibiotic or herbicide resistance genes are required because the GUS gene is both a selectable and a screenable marker gene at the same time. The use and release of selectable genes such as antibiotic resistance genes into the environment have been the target of concern among environmental authorities. While such concerns may prove unfounded, as suggested by Flavell et
al. (1992), they may nevertheless lead to governmental restrictions on the use of antibiotic resistance genes in transgenic plants, and it is therefore desirable to develop new selection methods which are independent of such genes. Here we have briefly outlined the principle for one such method.
Acknowledgment. The
authors would like to thank Ms. Eva Sogfird for skilful technical assistance.
References
Bevan M, Flavell RB, Chilton MD (1983) Nature 394:184-!87 Bowen BA (1993) In 'Transgenic Plants' Vol. 1. Eds.: Kung SD, Wu R. Academic Press pp. 89-146. De Block M, Botterman J, Vandewiele M, Docky J, Toen C, Gossete V, Movva NR, Thompson C, Van Montagu M, Leemans J (1987) EMBO J. 6:2513-2518 Fishman WH (1955) Advan. Enzymol. 16:361-409 Flavell RB, Dart E, Fuchs RL, Fraley RT (1992) Bio/Technology 10:141-144 Fraley RT, Rogers SG, Horsch RB (1986) CRC Critical Reviews in Plant Science 4: 1-45. Fraley RT, Rogers SG, Horsch RB, Sanders P, Flick J, Adams S, Bitmer M, Brand L, Fink C, Fray J, Galluppi G, Goldberg S, Woo S (1983) Proc. Natl. Acad. Sci. USA 80:4803-4806 Hille J, Verheggen F, Roelvink R Franssen H, Kammen AV, Zabel P (1986) Plant Mol. Biol. 7:171-176 Jefferson RA, Kavanagh TA, Bevan MW (1987) EMBO J. 6: 39013907 Murashige T, Skoog F (1962) Physiol. Plant. 15:473-497 Palmgren G, Mattson O, Okkels FT (1990) Biochem. Biophys. Acta 1049: 293- 297. Shah D M, Horsch RB, Klee HJ, Kishore GM, Winter JA, Tumer NE, Hironaka CM, Sanders PR, Gasser CS, Aykent S, Siegel NR, Rogers SG, Fraley RT (1986) Science 233:478-481 Stalker DM, McBride KF, Malyj LD (1988) Science 242:419-423 Streber WR, Wilhnitzer L (1989) Bio/Technology 7:811-816 Waldron C, Murphy EB, Roberts JL, Gustafson GD, Armour SL, Malcolm SK (1985) Plant Mol. Biol. 5:103-108
Plant Cell Reports (1996) 16:219-221
© Springer-Verlag1996
A novel principle for selection of transgenic plant cells: positive selection Morten Joersbo and Finn T. Okkels* Danisco Biotechnology, Langebrogade l, DK-1001 Copenhagen K, Denmark * Present address: Dansk Planteforaedling, Hojerupvej 31, DK-4660 Store Heddinge, Denmark Received 5 February 1996/Revised version received 8 May 1996 - Communicated by H. L6rz
Summary. A novel principle for selection of transgenic plant cells is presented. In contrast to traditional selection where the transgenic cells acquire the ability to survive on selective media while the non-transgenic cells are killed (negative selection), this selection method actively favours regeneration and growth of the transgenic cells while the non-transgenic cells are starved but not killed. Therefore, this selection strategy is termed 'positive selection'. The E. coli 13-glucuronidase gene was used as selectable (as well as screenable) gene and a glucuronide derivative of the cytokinin benzyladenine as selective agent which is inactive as cytokinin but, upon hydrolysis by GUS, active cytokinin is released stimulating the transformed cells to regenerate. Selection of Agrobacterium tumefaciens inoculated of tobacco leaf discs on benzyladenine N-3-glueuronide (7.5-15 mg/1) resulted in 1.7-2.9 fold higher transformation frequencies compared to kanamycin selection. A significant advantage of this selection procedure is the elimination of the need for herbicide and antibiotic resistance genes.
Abbreviations GUS - f3-glucuronidase, HPLC - High pressure liquid chromatography, NPT II - Neomycin phosphotransferase II, X-gluc - 5-bromo-4-chloro-3-indolylglucuronide.
Introduction For genetic transformation of plants, it is well known that only a minor fraction of the treated cells become transgenic while the majority of the cells remain untransforreed which is why they have to be eliminated by selection. The general strategy has been to introduce a selecCorrespondence to: M. Joersbo
table gene along with the gene(s) encoding the desired trait allowing the transformed cells to grow on selective media while the non-transformed cells are killed (Bowen 1993). The most widely used selectable gene is the neomycin phosphotransferase II (NPTII) gene (Fraley et al. 1986) which confers resistance to the aminoglycoside antibiotics kanamycin, neomycin and G-418 (Bevan et al. 1983). A number of other selective systems has been developed based on resistance to bleomycin (Hille et al. 1986), bromoxynil (Stalker et al. 1988), chloramphenicol (Fraley et al. 1983), 2,4-dichlorophenoxy-acetic acid (Streber and Willmitzer 1989), glyphosate (Shah et al. 1986), hygromycin (Waldron et al. 1985) or phosphinothricin (De Block et al. 1987). All of these compounds are antibiotics or herbicides. The corresponding resistance genes are in many cases not relevant to the desired transgenic trait and they may be undesirable after selection has been accomplished. Here we present a novel concept for selection of transgenic plant cells. The method exploits the fact that cytokinin must be added to plant explants in order to obtain optimal shoot regeneration rates. By adding cytokinin as an inactive glucuronide derivative, cells which have acquired the GUS gene by transformation are able to convert the cytokinin glucuronide to active eytokinin (Fig. 1) while untransformed cells are arrested in development. Because of the metabolic advantage the transgenic cells obtain, this is a positive selection strategy as opposed to traditional selection where the transgenic cells only become insensitive to the selective agent. In the positive selection system, the GUS gene serves the dual purpose of being both a selectable and screenable marker gene. Therefore, no herbicide or antibiotic resistance genes are needed.
220 Materials and methods Cytokinin glucuronide. Benzyladenine N-3-glucuronide was synthesized by Dr. R. Whenham, Apex Organics, Devon, UK (manuscript submitted) and was tested as a substrate for the E. coli GUS (Sigma G7896) by incubating a 1.0 mg/ml solution in 50 mM Naphosphate buffer pH 7.0 in the presence of GUS for 18 hours at 37°C. HPLC analysis (conditions modified according to Palmgren et al. 1990) showed a virtually complete removal of the benzyladenine N-3-glucuronide peak with concomitant production of a peak which cochromatographed with authentic benzyladenine. Transformation. Sterile shoot cultures of tobacco OVicotiana tabacum vat. 'Wisconsin 38') were maintained on MS-medium (Murashige and Skoog 1962). Leaf discs from the largest leaves (3-5 weeks old) were punched out and was transformed according to standard procedures using Agrobacterium tumefaciens strain LBA 4404 harbouring pBI121 containing both a GUS and NPTII gene (Jefferson et al. 1987). Selection. After co-culture for 3 days the leaf discs were transferred to selection medium (MS-medium) containing no cytokinin or kanamycin but various concentrations of benzyl-adenine N-3-glucuronide according to the experiment. In the controls, 1.0 mg/t benzyladenine and 300 mg/1 kanamycin sulfate were added instead ofbenzyladenine N-3-glucuronide. All selection media contained 800 rag/1 carbenicillin in order to eliminate the Agrobacteria. Shoot tips from regenerated shoots were tested for GUS activity by the X-gluc histological assay according to Jefferson et al. (1987). Only ieaf tips displaying strong blue reaction were evaluated as being positive. For each selection medium, 90-120 leaf discs were initiated in 3 or more independent experiments.
NH
CH2" ~
~----~l
N
Results
Incubation o f benzyladenine N-3-glucuronide with GUS resulted in total hydrolysis as evidenced by the release o f benzyladenine. This demonstrated that benzyladenine N3-glucuronide is substrate for GUS. Benzyladenine N-3-glucuronide was tested as a selective agent at various concentrations in the tobacco leaf disc transformation system and was found to give transformation frequencies which were superior to those obtained with kanamycin selection (Tab. 1). At 7.5 and 15 rag/1 benzyladenine N-3-glucuronide the transformation frequencies were 0.15 and 0.26 transformed shoot per explant, respectively, compared to 0.09 by kanamycin selection. Saccharo- 1,4-1actone is a potent inhibitor o f GUS (Fishman 1955). Addition of saccharo-l,4-1actone (at 1.92 g/l) reduced the transformation rate to 0.02 transformed shoots per explant indicating that the selection is dependent on the presence o f active GUS and that unhydrolysed benzyladenine N-3-glucuronide has only a very weak if any cytokinin activity. Saccharo-1,4-1actone up to at least 3.5 g/1 had no significant effect on the regeneration o f shoots from non-transgenic explants.
NH,, CH2
O
II
GUS
C~OH
lo ° H
HO
\
H
Glucuronic acid
( OH
Benzyladenine N-3-glucuronide
Benzyladenine
Fig. 1. Principle of positive selection. In cells transformed with a GUS gene the inactive cytokinin(benzyladenineN-3-glucuronide)is hydrolysedby GUS to active cytokinin (benzyladenine)which stimulates shoot regeneration and growth of the transgenic cells.
221 Benzyladenine N-3-glucuronide (mg/1)
Benzyladenine (mg/1)
Saccharo-l,4lactone (g/l)
Kanamycin sulfate (mg/1)
GUS+ shoots per leaf disc (+ SD)
7.5
0
0
0
0.15 (+ 0.11)
15.0
0
0
0
0.26 (_+ 0.14)
15.0
0
1.92
0
0.02 (_+ 0.01)
0
1.0
0
300
0.09 (_+ 0.04)
Tab. 1. Transformationfrequencies obtainedby positiveselection of tobacco shoots selected on media containing benzyladenine N-3-glucuronide in the presence or absence of saccharo-l,4-1actone, comparedto traditional selection on kanamycin (1.0 mg/1 benzyladenine and 300 mg/1 kanamycin sulfate). The transformation frequencies along with standard deviations (SD) were calculatedas the averagenumberof GUS-positiveshoots per leaf disc in 3 or more independent experiments. Discussion
We have presented a novel principle for selection oftransgenic plant cells, based on the fact that in vitro grown plant tissue cultures are auxotrophic in several aspects. In particular, exogenously applied cytokinin significantly enhances regeneration and growth. Due to the acquired ability to hydrolyse benzyladenine glucuronide, the transgenic cells harbouring the GUS gene were stimulated compared to the non-transgenic cells, hence the term 'positive selection'. Selection of transgenic plant cells by positive selection rather than negative selection offers significant advantages. Firstly, the transformation frequencies obtained by positive selection appear to be higher than those obtained with kanamycin selection. This could be related to the fact that during negative selection the majority of the cells in the explants die. Such dying cells may release toxic substances (such as phenolics) which in turn may impair regeneration of the transformed cells (Bowen 1993). In addition, dying cells may form a barrier between the medium and the transgenic cells preventing or slowing uptake of essential nutrients. The overall effect of necrotic tissue during selection is presumably reduced mitotic activity of the transformed cells resulting in less transgenic shoots emerging from the explants. When benzyladenine N-3-glucuronide was used as selective agent the explants maintained significantly higher viability during the selection procedure, compared to kanamycin selection. Secondly, no antibiotic or herbicide resistance genes are required because the GUS gene is both a selectable and a screenable marker gene at the same time. The use and release of selectable genes such as antibiotic resistance genes into the environment have been the target of concern among environmental authorities. While such concerns may prove unfounded, as suggested by Flavell et
al. (1992), they may nevertheless lead to governmental restrictions on the use of antibiotic resistance genes in transgenic plants, and it is therefore desirable to develop new selection methods which are independent of such genes. Here we have briefly outlined the principle for one such method.
Acknowledgment. The
authors would like to thank Ms. Eva Sogfird for skilful technical assistance.
References
Bevan M, Flavell RB, Chilton MD (1983) Nature 394:184-!87 Bowen BA (1993) In 'Transgenic Plants' Vol. 1. Eds.: Kung SD, Wu R. Academic Press pp. 89-146. De Block M, Botterman J, Vandewiele M, Docky J, Toen C, Gossete V, Movva NR, Thompson C, Van Montagu M, Leemans J (1987) EMBO J. 6:2513-2518 Fishman WH (1955) Advan. Enzymol. 16:361-409 Flavell RB, Dart E, Fuchs RL, Fraley RT (1992) Bio/Technology 10:141-144 Fraley RT, Rogers SG, Horsch RB (1986) CRC Critical Reviews in Plant Science 4: 1-45. Fraley RT, Rogers SG, Horsch RB, Sanders P, Flick J, Adams S, Bitmer M, Brand L, Fink C, Fray J, Galluppi G, Goldberg S, Woo S (1983) Proc. Natl. Acad. Sci. USA 80:4803-4806 Hille J, Verheggen F, Roelvink R Franssen H, Kammen AV, Zabel P (1986) Plant Mol. Biol. 7:171-176 Jefferson RA, Kavanagh TA, Bevan MW (1987) EMBO J. 6: 39013907 Murashige T, Skoog F (1962) Physiol. Plant. 15:473-497 Palmgren G, Mattson O, Okkels FT (1990) Biochem. Biophys. Acta 1049: 293- 297. Shah D M, Horsch RB, Klee HJ, Kishore GM, Winter JA, Tumer NE, Hironaka CM, Sanders PR, Gasser CS, Aykent S, Siegel NR, Rogers SG, Fraley RT (1986) Science 233:478-481 Stalker DM, McBride KF, Malyj LD (1988) Science 242:419-423 Streber WR, Wilhnitzer L (1989) Bio/Technology 7:811-816 Waldron C, Murphy EB, Roberts JL, Gustafson GD, Armour SL, Malcolm SK (1985) Plant Mol. Biol. 5:103-108
