PlantCell Reports
Plant Cell Reports (1994) 13:556 -560
9 Springer-Verlag1994
Genotypic 'and developmental regulation of transient expression of a reporter gene in soybean zygotic cotyledons Patricia J. Moore 1, ,, Allen J. Moore 2, and Glenn B. Collins 1
1 Department of Agronomy, University of Kentucky, Lexington, KY 40546-0091, USA 2 Department of Entomology, University of Kentucky, Lexington, KY 40546-0091, USA * Present address: Division of Natural Sciences and Mathematics, Transylvania University, 300 North Broadway, Lexington, KY 40508, USA Received 28 July 1993/Revisedversion received 24 March 1994 - Communicated by W. A. Parrott
Abstract Processes involved in transformation of regenerable soybean (Glycine max (L.) Merrill) immature zygotic cotyledons were studied by assaying the transient expression of the 13-giucuronidase gene driven by the 35S promoter and terminated at the 3' end by the soybean 7S storage protein gene. The plasmid containing the chimeric gene was delivered to the cotyledons via particle bombardment 900 PSI. Zygotic cotyledons from six soybean varieties were tested for transient expression of the 13-glucuronidase gene, The level of reporter gene expression differed between genotypes. The genotypes could be classified as high, fair or poor expressors. Cotyledons from different genotypes were then bombarded at 650, 900 or 1100 PSI. GUS expression varied among genotypes independently of the pressure of bombardment. Finally, the ability of cotyledons to express the reporter gene depended on the developmental stage of the seed from which it was excised with the younger stage being the least responsive. However, genotypic specific expression remained after controlling for developmental stage of the cotyledons. Abbreviations ANOVA, analysis of variance; GUS, 13-giucuronidase; PSI, pounds per square inch; TEUs, transient expression units; MSO3, MS media with 3% sucrose. Introduction
Current cultivars and breeding lines of soybean, Glycine max (L.) Merrill, developed through classical breeding methods, are characterized as having an extremely narrow genetic base (Christou et al. 1990). This limited gene pool restricts the introduction of new traits via traditional breeding techniques. Thus soybean is an ideal crop for the use of genetic engineering approaches
Correspondence to: P. J. Moore
for developing genotypes possessing desirable traits such as increased herbicide and disease resistance, drought tolerance and modification of oil and protein quality (Christou et al. 1990, Hinchee et al. 1988, Hildebrand et al. 1991). At present, however, genetic manipulation of soybean through recombinant DNA technology is severely limited by the lack of an efficiently coupled transformation and regeneration system for this species. One of the major impediments to the development of a transformation system for soybean has been the low efficiency of introducing foreign DNA into soybean cells. To date, Agrobacterium tumefaciens has been the most widely used vector for introducing DNA into dicotyledonous plant cells (Horsch et al. 1985, Hooykaas 1989). However, the use of Agrobaeterium to introduce DNA into soybean has not been highly successful as soybean is not generally considered a host for Agrobacterium infection (Matthysse and Guflitz 1982). While gall formation on soybean has been documented (Pederson et al. 1983, Hood et al. 1986), and generation of transgenic soybean through Agrobacterium-mediated gene transfer has been accomplished (Hinchee et al. 1988), the efficiency is low and soybean wound tissue only remains susceptible to Agrobacterium infection for a short period of time (Kudirka et al. 1986). Further, soybean susceptibility to Agrobacterium is highly genotype specific (Owens and Cress 1985, Bailey et al. 1993), limiting the ability to introduce novel genes directly into elite cultivars. Some researchers have turned to the use of accelerated microprojectiles for DNA delivery (Klein et al. 1987, Sato et al. 1993) in order to overcome the problems associated with Agrobacterium-mediated gene transfer. Recovery of transformed soybean plants has been reported using microprojectiles as the method of gene delivery (Christou et al. 1989, 1990, Finer and
557 McMullen 1991, McCabe et al. 1988). Since microprojectile delivery relies on physical parameters, it has been assumed that problems associated with the biological vector, Agrobacterium, are avoided and that a large number of explants and genotypes can be used as targets for transformation (Christou et al. 1990, Moore and Collins 1993). In order to test the hypothesis that biolistic transformation is genotype independent in soybean, we undertook a study to compare transient expression in zygotic cotyledons of various genotypos of soybean. Soybean can be regenerated from zygotic embryos as an initial explant. Particle bombardment of zygotic embryos has been used as a transformation system in rice (Li et al. 1993). Materials and Methods Plant materials Soybean plants were grown in a greenhouse as described in Parrott et al. (1988). Pods were surface-sterilized by immersion in 70% isopropyl alcohol for 30 see followed by 20% commercial chlorine bleach for 12 rain. Pods were rinsed 3 times in stea-Jle, distilled water, and immahlre seeds were removed. Cotyledons were excised from 4 to 6 m m seeds by removing the end ccntaining the embryonic axis and pushing the cotyledons out ofthe seed coat (Lazzeri et al. 1985). The cotyledons were placed abaxial side down on sterile filter paper on MSO3 medium camsisting of MS salts 0Viurashige and Skoog 1962), B5 vitamins (Gamborg et al. 1968), 3% sucrose, 0.2% Phytagel (SIGMA Chemical) as a solidifying agent, and adjusted to pI-I 5.8. Prior to explanting the cotyledous, each filter paper had born divided into six squares. A gmotype (Century, Clark 63, Fayette, 3103, McCall and Peking) was randomly assigned to each of the six positions, sxlch that each filter paper contained a cotyledon from all ge~aotypes use&
separate wells of a 96-well micrcCdter dish. The cotyledons were incubated for 16 hrs at 37 ~ C. Cotyledons were removed from each well and the number of TEUs (blue spots) were cotmted_ To determine in which cells GUS was being expressed, McCall zygotic cotyledons were bombarded at 900 PSI as described, The cotyledons were free-hand sectioned 2 days after bombardm~at prior to being placed into GUS histochemical assay buffe~.
Experimental Design and Statistical analysis We evaluated sources of the variation in GUS activity levels by performing a series of expeaq_mentscontrolling for different factors. In the first experimmt we tested for the effeas of genotype and DNA preparation on the number of transient expression units obseawed in a cotyledon. The six differe~at genotypes described above were tested in this experiment along with two separate preparatic~ls of DNA. The experiment was repeated twelve times. LogRransformed data were analyzed using a 2way ANOVA with D N A preparatiatts and soybean genotypes as factors. The second set of experiments examined the effect of g~otype and the pressure at which cotyledons were bombarded on TEUs. Iu this experiment, the sanae six genotypes were bombarded at 650, 900, or 1100 PSI. Six replicates were shot at each pressure. Log4_ransformed data were analyzed using a 2-way ANOVA with genotypes and PSI as factors. The third exp~ime~at examined the influence of develapmental stage ofthe cotyledons on TEUs. McCall cotyledclas were excised from seeds 3 man, 5 m m and 8 m m in length and bombarded with the GUS expression plasmid- These seed sizes are in the expansion stage of growth. A single cotyled~ from two seeds of each size was explanted onto a randomly assigned position on filter paper for each shot thus permitting six cotyledons to be shot simultaneously as above. This experiment was replicated 10 times. Due to differences in the surface area of the cx~tyledc~ b~we~a the three sizes of seeds, the number of TEUs per m m 2 was calculated- Log-transformed data were analyzed using a one-way A_NOVA with developmental stage as the factor. In order to eliminate potential differences in GUS expression due to disparity in the developmental stage of the cotyledons, zygotic cotyledons from seed above the apparent developmental threshold for GUS expressi~a (8 to 11 m m in length) from all genotypes but Clark were bombarded and assayed for GUS expression in a separate experiment. Eight replicates were bombarded. Log-transformed T.E.U.s .were analyzed by one-way ANOVA with genotype as the factor.
Particle bombardment The cotyledons were bombarded using the Biolistic PaIticle Delivery System 1000 fitted with a helium attachment (PDS-1000/He; Bio-Rad). The plasmid used for these studies contained a single 35S promoter from Cauliflower Mosaic virus driving the GUS coding region (Jefferson 1987) followed by the termination region from the ct subunit of the s o y b e ~ 78 storage protein (Sebastiani et al. 1990). This GUS expression cassette was cloned into a pUC19 plasmid. Mieroprojectiles were prepared by mixing 50 Ixl from a 60 mg/ml su~peusion of 1.1 p m diameter tungsten particles (M17; Bio-Rad) with 5 Ixl of a 1 Ixg,/Ixl solution ofplasmid D N A in TE. Spermidine was omitted from the DNA preparation because it has no effect on transient expression in this system (Moore, data not shown). This suapension was then mixed with 50 1~1 of 2.5 M Cat12. The resulting suspension of microprojectiles was washed one time with 250 ~tl of absolute ahanol and then resuspended in 60 ixl of absolute ethanol. 7.5 pl of the mieroprojeaile suspension was immediately loaded onto each macroprojectile disk and allowed to air dry. For each bombardment, the lifter paper on which the cotyledons were placed was removed from the MSO3 medium and placed in a dry, sterile p ari dish. This dish was placed 95 m m below the rupture disk and bombarded at 900 PSI, except in one experiment where material was bombarded at either 650, 900, or 1100 PSI. In the experiments where more than one genotype was evaluated, we bombarded all of the genotypes in a single shot. Within a shot a single cotyledon from each genotype was placed in one of six randomly assigned positions on a piece of sterile filter paper. Following bombardment each filter paper was removed from the dry dish and placed back onto MSO3 medium.
GUS histochemical assay Two days after bombardment, individual cotyledons were removed from the filter papers and placed into 250 Ixl of GUS histochemieal assay (50 ~ sodium phosphate buffer, pH 7.0, 10 m M EDTA, 0.5 m M ferrocyanide, 0.5 m M ferrieyanide, 0.05% Triton X-100) buffer in
Results
The genotype of the cotyledons affects the number of transient expression units. When l~-glucuronidase (GUS) activity was assayed histochemically two days post bombardment (Fig. 1), the mean number of transient expression units differed significantly between the six genotypes tested (Fig. 2). We found no significant differences among the two DNA preparations (F = 0.449, df = 1,60, p>0.5). However, there were highly significant differences among genotypes in the expression of GUS (F = 6.471, df = 5,60, p < 0.001). The differences anlong genotypes did not depend on the DNA preparation used (interaction term; F = 1.382, df = 5,60, p > 0.20). Tukey's,HSD post-hoe comparison test grouped the genotypes into three categories: McCall as a good expressor; Century, Clark 63 and J103 as fair expressors; and Fayette and Peking as poor expressors.
558
The PSI at which cotyledons are bombarded as well as the genolype influences transient expression of GUS, but the genotype effect is not dependent on the PSI effect.
The developmental stage of the zygotic cotyledons influences their ability to express GUS following bombardment. Developmental stage had a significant
The mean number of transient expression units for each genotype following bombardment at 650 PSI, 900 PSI and 1100 PSI are shown in Figure 3. The pressure at which cotyledons were bombarded has a large effect on the transient expression of GUS (F = 10.877, df = 2,90, p 0.90).
effect on expression of GUS (Fig. 4; F -- 5.902, df = 2,27, p
Plant Cell Reports (1994) 13:556 -560
9 Springer-Verlag1994
Genotypic 'and developmental regulation of transient expression of a reporter gene in soybean zygotic cotyledons Patricia J. Moore 1, ,, Allen J. Moore 2, and Glenn B. Collins 1
1 Department of Agronomy, University of Kentucky, Lexington, KY 40546-0091, USA 2 Department of Entomology, University of Kentucky, Lexington, KY 40546-0091, USA * Present address: Division of Natural Sciences and Mathematics, Transylvania University, 300 North Broadway, Lexington, KY 40508, USA Received 28 July 1993/Revisedversion received 24 March 1994 - Communicated by W. A. Parrott
Abstract Processes involved in transformation of regenerable soybean (Glycine max (L.) Merrill) immature zygotic cotyledons were studied by assaying the transient expression of the 13-giucuronidase gene driven by the 35S promoter and terminated at the 3' end by the soybean 7S storage protein gene. The plasmid containing the chimeric gene was delivered to the cotyledons via particle bombardment 900 PSI. Zygotic cotyledons from six soybean varieties were tested for transient expression of the 13-glucuronidase gene, The level of reporter gene expression differed between genotypes. The genotypes could be classified as high, fair or poor expressors. Cotyledons from different genotypes were then bombarded at 650, 900 or 1100 PSI. GUS expression varied among genotypes independently of the pressure of bombardment. Finally, the ability of cotyledons to express the reporter gene depended on the developmental stage of the seed from which it was excised with the younger stage being the least responsive. However, genotypic specific expression remained after controlling for developmental stage of the cotyledons. Abbreviations ANOVA, analysis of variance; GUS, 13-giucuronidase; PSI, pounds per square inch; TEUs, transient expression units; MSO3, MS media with 3% sucrose. Introduction
Current cultivars and breeding lines of soybean, Glycine max (L.) Merrill, developed through classical breeding methods, are characterized as having an extremely narrow genetic base (Christou et al. 1990). This limited gene pool restricts the introduction of new traits via traditional breeding techniques. Thus soybean is an ideal crop for the use of genetic engineering approaches
Correspondence to: P. J. Moore
for developing genotypes possessing desirable traits such as increased herbicide and disease resistance, drought tolerance and modification of oil and protein quality (Christou et al. 1990, Hinchee et al. 1988, Hildebrand et al. 1991). At present, however, genetic manipulation of soybean through recombinant DNA technology is severely limited by the lack of an efficiently coupled transformation and regeneration system for this species. One of the major impediments to the development of a transformation system for soybean has been the low efficiency of introducing foreign DNA into soybean cells. To date, Agrobacterium tumefaciens has been the most widely used vector for introducing DNA into dicotyledonous plant cells (Horsch et al. 1985, Hooykaas 1989). However, the use of Agrobaeterium to introduce DNA into soybean has not been highly successful as soybean is not generally considered a host for Agrobacterium infection (Matthysse and Guflitz 1982). While gall formation on soybean has been documented (Pederson et al. 1983, Hood et al. 1986), and generation of transgenic soybean through Agrobacterium-mediated gene transfer has been accomplished (Hinchee et al. 1988), the efficiency is low and soybean wound tissue only remains susceptible to Agrobacterium infection for a short period of time (Kudirka et al. 1986). Further, soybean susceptibility to Agrobacterium is highly genotype specific (Owens and Cress 1985, Bailey et al. 1993), limiting the ability to introduce novel genes directly into elite cultivars. Some researchers have turned to the use of accelerated microprojectiles for DNA delivery (Klein et al. 1987, Sato et al. 1993) in order to overcome the problems associated with Agrobacterium-mediated gene transfer. Recovery of transformed soybean plants has been reported using microprojectiles as the method of gene delivery (Christou et al. 1989, 1990, Finer and
557 McMullen 1991, McCabe et al. 1988). Since microprojectile delivery relies on physical parameters, it has been assumed that problems associated with the biological vector, Agrobacterium, are avoided and that a large number of explants and genotypes can be used as targets for transformation (Christou et al. 1990, Moore and Collins 1993). In order to test the hypothesis that biolistic transformation is genotype independent in soybean, we undertook a study to compare transient expression in zygotic cotyledons of various genotypos of soybean. Soybean can be regenerated from zygotic embryos as an initial explant. Particle bombardment of zygotic embryos has been used as a transformation system in rice (Li et al. 1993). Materials and Methods Plant materials Soybean plants were grown in a greenhouse as described in Parrott et al. (1988). Pods were surface-sterilized by immersion in 70% isopropyl alcohol for 30 see followed by 20% commercial chlorine bleach for 12 rain. Pods were rinsed 3 times in stea-Jle, distilled water, and immahlre seeds were removed. Cotyledons were excised from 4 to 6 m m seeds by removing the end ccntaining the embryonic axis and pushing the cotyledons out ofthe seed coat (Lazzeri et al. 1985). The cotyledons were placed abaxial side down on sterile filter paper on MSO3 medium camsisting of MS salts 0Viurashige and Skoog 1962), B5 vitamins (Gamborg et al. 1968), 3% sucrose, 0.2% Phytagel (SIGMA Chemical) as a solidifying agent, and adjusted to pI-I 5.8. Prior to explanting the cotyledous, each filter paper had born divided into six squares. A gmotype (Century, Clark 63, Fayette, 3103, McCall and Peking) was randomly assigned to each of the six positions, sxlch that each filter paper contained a cotyledon from all ge~aotypes use&
separate wells of a 96-well micrcCdter dish. The cotyledons were incubated for 16 hrs at 37 ~ C. Cotyledons were removed from each well and the number of TEUs (blue spots) were cotmted_ To determine in which cells GUS was being expressed, McCall zygotic cotyledons were bombarded at 900 PSI as described, The cotyledons were free-hand sectioned 2 days after bombardm~at prior to being placed into GUS histochemical assay buffe~.
Experimental Design and Statistical analysis We evaluated sources of the variation in GUS activity levels by performing a series of expeaq_mentscontrolling for different factors. In the first experimmt we tested for the effeas of genotype and DNA preparation on the number of transient expression units obseawed in a cotyledon. The six differe~at genotypes described above were tested in this experiment along with two separate preparatic~ls of DNA. The experiment was repeated twelve times. LogRransformed data were analyzed using a 2way ANOVA with D N A preparatiatts and soybean genotypes as factors. The second set of experiments examined the effect of g~otype and the pressure at which cotyledons were bombarded on TEUs. Iu this experiment, the sanae six genotypes were bombarded at 650, 900, or 1100 PSI. Six replicates were shot at each pressure. Log4_ransformed data were analyzed using a 2-way ANOVA with genotypes and PSI as factors. The third exp~ime~at examined the influence of develapmental stage ofthe cotyledons on TEUs. McCall cotyledclas were excised from seeds 3 man, 5 m m and 8 m m in length and bombarded with the GUS expression plasmid- These seed sizes are in the expansion stage of growth. A single cotyled~ from two seeds of each size was explanted onto a randomly assigned position on filter paper for each shot thus permitting six cotyledons to be shot simultaneously as above. This experiment was replicated 10 times. Due to differences in the surface area of the cx~tyledc~ b~we~a the three sizes of seeds, the number of TEUs per m m 2 was calculated- Log-transformed data were analyzed using a one-way A_NOVA with developmental stage as the factor. In order to eliminate potential differences in GUS expression due to disparity in the developmental stage of the cotyledons, zygotic cotyledons from seed above the apparent developmental threshold for GUS expressi~a (8 to 11 m m in length) from all genotypes but Clark were bombarded and assayed for GUS expression in a separate experiment. Eight replicates were bombarded. Log-transformed T.E.U.s .were analyzed by one-way ANOVA with genotype as the factor.
Particle bombardment The cotyledons were bombarded using the Biolistic PaIticle Delivery System 1000 fitted with a helium attachment (PDS-1000/He; Bio-Rad). The plasmid used for these studies contained a single 35S promoter from Cauliflower Mosaic virus driving the GUS coding region (Jefferson 1987) followed by the termination region from the ct subunit of the s o y b e ~ 78 storage protein (Sebastiani et al. 1990). This GUS expression cassette was cloned into a pUC19 plasmid. Mieroprojectiles were prepared by mixing 50 Ixl from a 60 mg/ml su~peusion of 1.1 p m diameter tungsten particles (M17; Bio-Rad) with 5 Ixl of a 1 Ixg,/Ixl solution ofplasmid D N A in TE. Spermidine was omitted from the DNA preparation because it has no effect on transient expression in this system (Moore, data not shown). This suapension was then mixed with 50 1~1 of 2.5 M Cat12. The resulting suspension of microprojectiles was washed one time with 250 ~tl of absolute ahanol and then resuspended in 60 ixl of absolute ethanol. 7.5 pl of the mieroprojeaile suspension was immediately loaded onto each macroprojectile disk and allowed to air dry. For each bombardment, the lifter paper on which the cotyledons were placed was removed from the MSO3 medium and placed in a dry, sterile p ari dish. This dish was placed 95 m m below the rupture disk and bombarded at 900 PSI, except in one experiment where material was bombarded at either 650, 900, or 1100 PSI. In the experiments where more than one genotype was evaluated, we bombarded all of the genotypes in a single shot. Within a shot a single cotyledon from each genotype was placed in one of six randomly assigned positions on a piece of sterile filter paper. Following bombardment each filter paper was removed from the dry dish and placed back onto MSO3 medium.
GUS histochemical assay Two days after bombardment, individual cotyledons were removed from the filter papers and placed into 250 Ixl of GUS histochemieal assay (50 ~ sodium phosphate buffer, pH 7.0, 10 m M EDTA, 0.5 m M ferrocyanide, 0.5 m M ferrieyanide, 0.05% Triton X-100) buffer in
Results
The genotype of the cotyledons affects the number of transient expression units. When l~-glucuronidase (GUS) activity was assayed histochemically two days post bombardment (Fig. 1), the mean number of transient expression units differed significantly between the six genotypes tested (Fig. 2). We found no significant differences among the two DNA preparations (F = 0.449, df = 1,60, p>0.5). However, there were highly significant differences among genotypes in the expression of GUS (F = 6.471, df = 5,60, p < 0.001). The differences anlong genotypes did not depend on the DNA preparation used (interaction term; F = 1.382, df = 5,60, p > 0.20). Tukey's,HSD post-hoe comparison test grouped the genotypes into three categories: McCall as a good expressor; Century, Clark 63 and J103 as fair expressors; and Fayette and Peking as poor expressors.
558
The PSI at which cotyledons are bombarded as well as the genolype influences transient expression of GUS, but the genotype effect is not dependent on the PSI effect.
The developmental stage of the zygotic cotyledons influences their ability to express GUS following bombardment. Developmental stage had a significant
The mean number of transient expression units for each genotype following bombardment at 650 PSI, 900 PSI and 1100 PSI are shown in Figure 3. The pressure at which cotyledons were bombarded has a large effect on the transient expression of GUS (F = 10.877, df = 2,90, p 0.90).
effect on expression of GUS (Fig. 4; F -- 5.902, df = 2,27, p
