The development of gene transfer techniques for plant species is of great interest and value to plant breeders because it can be used for the rapid transfer of beneficial genetic traits to plants. In particular, the development of gene transfer techniques for leguminous plants is of commercial interest because they facilitate the development of new cultivars with improved disease resistance, tolerance to specific herbicides and increased nutritional value. Recombinant DNA techniques are being used to transfer foreign genes into agronomically important crops and to study the expression and regulation of genes in plant systems. The meaningful application of these techniques is dependent upon efficient transfer of genes into cells capable of regenerating into fertile plants. Numerous methods have been developed for transferring genes into plant tissues including Agrobacterium-mediated transfer, direct DNA uptake, microinjec- tion, high-velocity microprojectiles and electroporation.

Agrobacterium-mediated gene transfers are by far the most widely used gene transfer techniques. U.S. Patent Number 4,459,355 issued July 10, 1984 to Cello et al. and U.S. Patent Number 4,940,838 issued July 10, 1990 to Schilperoot et al, both incorporated herein by reference, describes the background of the development and use Agrobacterium-mediated gene transfers. Unfortunately, the use of Agrobacterium strains for gene transfers can be limited. Although dicotyledonous species such as leguminous plants are susceptible to Agrobacterium infections, its use for transformation is limited due to the lack of efficient regeneration procedures for transformed tissues. U.S. Patent Number 4,945,050 ('050 patent) issued July 31, 1990 to Sanford et al., which is incorporated herein by reference, provides a review of the mechanisms and limitations of direct DNA uptake, cell fusions and microinjection as methods of introducing foreign genes into cells.

The '050 patent relates to a method of introducing foreign substances into living cells using microprojectiles. While the '050 patent teaches a method of transforming many cells simultaneously, problems associated with microprojectile transformation include: the high mortality of transformed cells; the complexity of equipment needed; and, the difficulties in operating the equipment to efficiently perform the method.

The present invention relates to transgenic soybean plants. Soybean species have been transformed at low frequency using the Agrobacterium or by particle bombardment method described above. Electroporation, one of the direct DNA transfer techniques, has been used to

stably transform a number of plants. Soybean transformation using electroporation and stable integration of genes in the calli have been reported, but efforts to regenerate plants are unsuccessful. Recently Wei and Xu (Plant Cell Reports 7:348-351 (1988)) reported plant regeneration from soybean protoplasts. However, the efficiency reported is low and did not use commercially important strains that are transformed by electroporation.

The present invention relates to an improved method of producing transgenic soybean plants by transforming soybean plant cells by electroporation and regenerating the transformed cells into mature soybean plants. According to the present invention, cells of commercially important soybean varieties can be transformed by electroporation in a stable and highly efficient manner to introduce important genetic material and the resulting transformed cells can be regenerated at high efficiency to mature soybean plants. Using the method of the present invention, many cells can be transformed simultaneously using relatively simple techniques requiring relatively simple equipment resulting in a high survivability of transformed cells which can be regenerated into mature transgenic soybean plants at high efficiencies. INFORMATION DISCLOSURE

Wei and Xu, Plant Cell Reports 7:348-351 (1988) disclose a method of regenerating soybean plant protoplasts into mature plants. The method described in Wei could not be replicated using commercial varieties as starting materials and the efficiency of regeneration reported by Wei and Xu is significantly lower than the efficiency achieved practicing the method according to the present invention which uses transformed protoplasts derived from commercially important strains.

Christou, P. et al., Proc. Natl. Acad. Sci. USA 84:3962-3966 (1987) have reported transforming linked genes in soybean protoplasts using electroporation. Differences between the procedure used and the present invention include the heat shock step added in the present invention. This difference results in higher transformation frequency by the present invention. Furthermore, the transformed cells reported in Christou are not regenerated into mature plants. Fromm, M., et al., Proc Natl Acad Sci USA 82:5824-5828 (1985) report expression of genes transferred into monocot and dicot plant cells by electroporation.

SUMMARY OF THE INVENTION The present invention relates to a method of generating transgenic soybean plants comprising the steps of: preparing protoplasts from soybean cotyledons; inserting foreign DNA into the prepared protoplasts by electroporation, the foreign DNA comprising a useful gene; culturing the electroplated protoplasts in medium to induce cell growth, colony formation and calli generation; and regenerating plants from calli. DETAILED DESCRIPTION OF THE PRESENT INVENTION

According to the present invention, electroporation mediated DNA transfer into

protoplasts prepared from immature cotyledons is a highly efficient transformation method for soybean. Previous reports on soybean transformation showed the presence of chimeric sectors in the transgenic plants produced by Agrobacterium and particle acceleration methods. This may be either due to cross protection of non-transformed tissues by transformed ones or multiple transformation events within a cell population that gives rise to callus or plants.

However, the protoplast-electroporation system according to the present invention allows for the selection of the transformed single cells which can divide and give rise to regenerable callus forming transgenic shoots. Based on the total number of microcalli formed with and without selection, a relative transformation frequency of 21.4-23.0% can be obtained, corresponding to an absolute transformation frequency of 5.7-6.8 x 10 . The transformation frequency achieved for the soybean variety Glycine max is higher than that reported earlier for other protoplast culture systems. Only with two other plant systems, N. tabacum and Oryza sativa have higher transformation frequencies (up to 1 x 10 ]) been achieved by using a combination of PEG and electroporation or by electroporation only. Plating and selection methods along with several other factors are important criteria in the development of a stable transformation and regeneration system. In one embodiment of the present invention, an agarose-embedding system is employed to immobilize the transformed cells and therefore essentially eliminate the possibility of multiple recovery of single events or recovery of 'false positives' due to inadequate selection pressure. Another element of the present invention is use of a selection marker which confers sufficient sensitivity to transformed cells carrying the chimeric gene construction. In one embodiment according to the present invention, hygromycin is employed as the selection agent. Hygromycin as a selection agent has been used in other systems, i.e. Arbidopsis, orchardgrase and rise. A transgene construct according to the present invention comprises a desired gene which confers a desired trait on the transgenic soybean plants which contain it. The present invention embodies the successful transformation of soybean protoplasts via electroporation and recovery of transgenic shoots. Additionally, a plant regeneration system from protoplast derived transformed calli is disclosed. Thus, it is possible to introduce, in a highly efficient manner, the stable inheritance and expression of genes, including those of agronomic interest, into soybean protoplasts at the whole plant level. To practice the present invention, plasmids which contain the genes to be introduced are constructed. Plasmids must contain the necessary genetic elements for expression in plant cells. These elements include an operably linked promoter and polyadenylation addition signal. In addition to the desired gene, the construct optionally contains a gene which encodes a selectable marker. The techniques to produce a gene construct useful in the present invention are well known to those having

ordinary skill in the art and the starting materials useful to practice the present invention are readily available.

Immature pods are collected from 60-80 day old soybean plants and surface sterilized. Cotyledons, 2-7 x 2-3 mm size are cut transversely into 0.5-2 mm thick sections, preferably 1-2 mm thick sections, and preplasmolyzed in CPW 13M. Preplasmolysis of the tissue in CPW 13M makes the plasma membrane shrink away from the cell wall, thus leaving space for enzyme to enter and act; the enzymatic treatment thereby becomes more effective. Washing the tissue helps in removing the damaged tissue thus the ultimate protoplast preparation is cleaner. The tissue is incubated in enzyme solution for about 4-6 hr in the dark. The enzyme solution consists of 1.5% (w/v) Cellulase "Onozuka" RIO, 0.2% Pectolyase Y23 dissolved in CPW 9M (pH 5.8). Among the several enzyme combinations tested for obtaining high protoplast yield and viability, the combination of Cellulase "Onozuka" R10 (1.5%) and Pectolyase Y23 (0.2%) has been found to be the optimum, because the viability of protoplasts was highest (about 85- 93%), and the yield adequate (5-6 x 10 ⁶ per gm fresh wt.). Use of Pectolyase Y23 seems to be necessary to obtain high yields of viable protoplasts. Protoplasts vary in size from 20-35 μm in diameter. Protoplasts are rarely released within the first two hours and incubation longer than six hours reduces the viability of protoplasts greatly. The most suitable duration of incubation is 4-6 hr. Enzyme combination used for preparing protoplasts is somewhat critical. Higher concentration of enzyme and longer incubation period are usually deleterious to protoplast preparations and may affect plating efficiency and plant regeneration. A much lower concentration of cellulose (1.5%) and use of Pectolyase Y23, in particular seems to be helpful in isolating protoplasts and in a relatively much shorter 4-6 hr compared to 18 hr time period. Cotyledon age and size are important factors influencing protoplast yield and viability. Optimum protoplast yield and viability is obtained from cotyledons of 3-4 x 2 mm in size. Protoplast release is poor from cotyledons of very large size (6-7 x 3 mm).

The released protoplasts are filtered through sieves and pelleted by centrifugation. At least 43 μm sieve is used. Preferably an additional step is to filter with a 74 μm sieve also because it makes the process of removing debris more efficient. The purpose of centrifugation is to get a pellet of living, uniform protoplasts. Protoplasts are washed twice by resuspending in either CPW 9M or KP8 medium. The washed protoplasts are then purified by floating over 23% (w/v) sucrose overlaid with 2.0 ml of KP8 medium and centrifuged. Protoplasts can be gently removed with a Pasteur pipette from the interface. Soybean protoplasts burst (3-7% of the total population, as assessed by counting with haemocytometer) during washing when CPW 9M is used. This can be avoided by using KP8 medium for washing and purification. When protoplasts are floated on sucrose (23%) there is no deleterious effect on viability. However, when Ficoll (13%) is used, 100% of the protoplasts burst within a few hours. Sucrose

concentration can be varied from 21-23%, without any major effect, washing after flotation is necessary to remove sucrose otherwise it can change the osmoticum and have deleterious effect on protoplast preparation.

Protoplasts are resuspended at the density of 1-2 x 10 ⁶ /ml in electroporation buffer (10 mM HEPES (pH 7.2), 150 mM NaCl, 5 mM CaCl ₂ 2H ₂ O and 0.2 M nannitol). A variety of buffers work including HBM and HBS which are the same as the buffer described above except with 1 mM Hepes and 29 mM Hepes, respectively. In addition, KP8 and 14 mM CaCl ₂ 7H ₂ O and sodium phosphate buffer also work.

One ml of protoplast suspension is chilled briefly in ice water and then heat-shocked at 45°C for 5 min prior to addition of 20μg/ml supercoiled plasmid DNA. Chilling is required to prevent resynthesis of the cell wall. Heat shock at 45°C for 5 min is useful as it presumably helps in DNA uptake by affecting the overall membrane permeability and protects the incoming DNA by switching off normal protein synthesis and inducing the synthesis of heat shock proteins, ultimately increases the transformation efficiency. Usually about 20 μg of DNA is added to 1 ml of protoplast suspension. While 10 to

100 μg also work, 20 μg per 1 ml of protoplast suspension is optimum. Subsequently, polyethylene glycol (20-30 PEG, preferably 28% - PEG, 6000 MW; 200 μl of 28% stock) dissolved in KP8 medium containing 120 mM MgCl ₂ is mixed with the protoplast:DNA suspension. Addition of 200 μl of 28% PEG is helpful, as it promotes the association of DNA with the membrane. Using lower PEG concentration, uptake of DNA is less; however, when used at higher concentration PEG causes fusion of protoplasts.

The samples are transferred to precooled, presterilized electroporation vessels fitted with electrodes. An electric field of 500 V/cm is applied by a single discharge of a 1000 μF capacitor that is precharged with an electrophoresis power supply. Aliquots electroporated without DNA and aliqouts treated with plasmid DNA without electroporation serve as controls. Heat shock treatment given to freshly isolated protoplasts prior to electroporation increases the percentage of dividing protoplasts observed by day 14. The stimulatory effect of heat shock on protoplast division is reflected in the increased number of colonies in control cultures and of the resistant colonies which develop following exposure of protoplast derived cells to selection medium.

Following electroporation, protoplasts are cultured for about 2 weeks without selection if selection is to be performed. Protoplast density is adjusted to 2.5 x 10 ⁵ protoplasts/ml in KP8 medium containing 2% Ficoll (Type 400-DL), 40 mM MES. Use of 2% ficoll (type 400 DL) in the liquid culture medium prevents the protoplasts from adhering to the surface of the petri dish. Otherwise along with the dead protoplasts the dividing ones also settle down and the toxic substances released by the dead ones may affect the growth of dividing colonies. Use of

40 mM MES buffer in the liquid medium helps in stabilizing the pH of the liquid medium (which usually drops down from 5.8 to 3.9) in the first 7 days of culture. During culture the osmolarity of the medium is progressively reduced by addition of a 1:1 mixture of KP8:K8 medium at day 7 and addition of an equal volume of a 1:2 mixture of KP8:K8 at day 10. Regular dilution of KP8 culture medium with K8 medium, helps in gradually lowering the osmoticum, which is again replaced with MSB medium gradually to provide various different additives for further growth of dividing cells and colonies.

If no selection procedure is to be performed, culturing in KP8:K8 medium is continued for about 5-6 weeks after electroporation. The proportion of K8 to KP8 is gradually increased. If a selection marker is included, selection procedures can be performed. As used herein the term "selection medium" refers to medium supplemented with compounds that are lethal to all cells except those having the protein produced by the expression of the selection marker. On day 16, the protoplast derived transformed cells, at the 2-4 cell stage, are selected by applying either liquid K8/selection medium, i.e., K8 medium supplemented with compounds which are lethal to all cells except those transformed with the selection marker, or the cells are resuspended in an equal volume of 1.2% Sea-plaque agarose (LMT, FMC Corp., ME, USA) gelled K8/selection medium. Solidified agarose cultures are cut into small beads and submerged in liquid K8/selection medium. In liquid medium, the influence of initial protoplast density reveals that at densities lower than 2.5 x lO^ml, cell budding can be frequently observed. Plating efficiencies of 38-50% can be observed after 7 days in culture. In parallel, the pH value drops 3.9-4.1 immediately, and as a result, frequency of cell division decreases dramatically and budding can be observed in nearly all the dividing protoplasts. However, gradual reduction in the osmolarity will stimulate development of cell colonies. When protoplasts are plated in agarose bead sectors and droplets at 2.5 x lO^/ml density, protoplast division is stimulated resulting in a plating efficiency of 52-58% . The use of 1.2% agarose reduces protoplast lysis during the initial stage of culture and further increases the plating efficiency up to 55-60% . The use of agarose bead culture technique minimizes pigment production and release of phenolic compounds and also offers the possibilities of removing deleterious compounds, easily stabilizing the pH and osmotic pressure, since the medium can be changed readily.

On selection medium the resistant colonies continue to grow and reach a size of 100- 200 μm or more after 4 weeks in culture. During selection, visible colonies develop in agarose beads and also in the surrounding liquid medium (released from the beads). Transformed colonies are also recovered when selection medium is used only in the liquid medium indicating that both the selection methods can yield large numbers of resistant colonies. However, the advantage of embedding the protoplasts in an agarose bead over the liquid selection method is

that the culture medium can be replaced without disturbing the developing colonies. The resistant colonies are easily distinguishable microscopically after about 4 weeks of selection (6 weeks after electroporation) and can be scored visually after about 6 weeks. Gradual dilution of KP8 medium with K8 using this time table is important, as it helps in gradually lowering the osmoticum. By using the agarose bead technique which immobilizes the transformants, replacement of liquid culture medium at regular intervals not only provides fresh culture medium (changing liquid is relatively easy) but also helps in removing the phenolic substances released by the dead cells. The use of 1.2% agarose instead of 0.6% agarose gives slightly higher plating efficiency, but the main purpose of agarose immobilization is to prevent protoplast lysis during the initial culture period.

Transformation frequency can determined by counting the number of colonies in the agarose blocks. The -native transformation frequency (RTF) is described as the ratio between the number of resistant colonies in the selected cultures and the developing colonies in the unselected cultures. The absolute transformation frequency (ATF) is calculated on the basis of number of resistant colonies produced after 8 weeks from the initial number of protoplasts plated after electroporation.

After about 5-6 weeks of culture, microcalli about 1-2 mm in size can be observed. The microcalli are transferred onto MSB medium supplemented with 0.5 mg mg/1 each of 2,4- D, BA, Kn, 500 mg/1 CH, 3% sucrose and 0.6% agar (with or without millipore 0.45 mm pore size filters). Where the colonies grow (to 3-5 mm) within 3-4 weeks and proliferate as light green calli. When grown on millipore filter paper the calli can grow faster and are more compact than the calli that are grown directly on culture medium. When the fast growing green calli are selectively transferred onto fresh medium of the same composition, green compact calli form after 3-4 subcultures. Following regular subculturing the dark green callus pieces with nodular structures are selectively transferred to MSB medium with various combinations of auxins and cytokinins for shoot formation. Such as MSB + BA + Kn + Z + NAA = CA ₃ + CH medium containing asparagine and glutamine. After repeated subculture, small shoots are formed but these shoots do not elongate on this medium. When transferred to MSB medium containing 3% sucrose and 0.5 mg/1 each of BA, Kn, ZT, 0.1 mg/1 NAA and 500 mg/1 CH and incubated under high light intensity (2,000 bu), in 2-3 weeks dark green nodular structures appear on the surface and periphery of the calli. In one experiment an average of 15 ± 3 green shoots developed from these meristematic nodules in 21.2% of the calli after 3-4 subcultures of 14 days each. Addition of 50 mg/1 each of glutamine and asparagine and 0.2 mg/1 GA ₃ further enhanced the number of shoots to 25 + _. 3 and the frequency up to 29.0% .

For further elongation shoots are transferred in 1/2 strength MS minerals containing 1.0

gm/1 KNO ₃ , 0.01 mg/1 TH with or without 0.5-1.0 mg/1 GA ₃ . Shoots (0.5-1.0 cm) are transferred to several different media for elongation. Elongation rate is significantly affected by growing the shoots on 1/2 strength MS minerals containing 1.0 gm/1 KNO3, 0.01 mg/1 TH and 1.0 mg/1 GA ₃ . In two or three weeks approximately 60% of the cultures produce 2-3 cm long shoots.

For rooting 2-3 cm long shoots are cultured in liquid or agar solidified 1/2 strength MS medium with 1 % sucrose and 0.2 mg/1 IBA or 0.5 mg/1 NAA and the cultures are initially kept in the dark for 10-12 days. Once root induction is observed, the plantlets are transferred to Magenta boxes containing autoclaved vermiculite and regularly watered with Hoagland's solution. The boxes are kept under high light and humidity conditions. When the plants attain a height of 6-8 cm and developed two to three new trifoliate leaves they are transferred to 3:1 soil and vermiculite mixture in bigger glass bottles with loose caps to lower the humidity. Later, the lids are removed from these bottles to expose the plants to culture room conditions. Finally, the plants are transferred to soil and are kept in the greenhouse. Protoplasts are isolated several times from the immature cotyledons obtained from greenhouse as well as field grown plants and from each experiment plating efficiencies in excess of 60% are obtained. Regenerable callus with multiple shoots and plants are recovered from most of the experiments. The calli seem to retain regeneration potential even after several months. In general, it took 4-5 months to regenerate plants from isolated protoplasts including rooting of the shoots.

Protoplast derived resistant calli are usually subcultured once every 2 weeks on MSB medium containing 0.5 mg/1 each of 2,4-D, BA, Kn and 500 mg/1 CH. After 4 subcultures, calli started becoming nodular. Selective subculturing of the nodular calli on MSB medium supplemented with 0.5 mg ^"1 each of BA, KN, ZT, 0.1 mg/1 of NAA, 0.2 mg/1 of GA ₃ and 50 mg/1 of asparagine and glutamine, triggered the differentiation of green buds and leafy shoots after 4-6 or more subcultures. Of 440 calli derived from 34 independent transformed clones maintained on regeneration medium, 35 calli regenerated 5-20 mm long shoots (10-12 per calli) after 6-8 subcultures, giving a regeneration frequency of approximately 8.0%. Leaf samples assayed indicate expression of the introduced gene. EXAMPLE 1

A reproducible plant regeneration method from immature cotyledon protoplasts of a commercial Glycine max genotype has been established. It is used in combination with high efficiency stable transformation of protoplast derived soybean cells via electroporation using a chimeric gene encoding hygromycin resistance and /3-glucuronidase and subsequent recovery of transformed shoots to generate transgenic soybean plants.

Stable transformation of soybean (Glycine max (L.) Merr.) immature cotyledon

protoplasts is achieved following electroporation with plasmid DNA carrying a chimeric gene for j8-glucuronidase (GUS) and hygromycin phosphotransferase (Hpt) under the control of the cauliflower mosaic virus (CaMV) 35S promoter. Transformed colonies are stringently selected by growing 15-day old protoplast derived cells in the presence of 40 μg/ml of hygromycin-B for 6 weeks. Over 93% of the resistant cells and colonies exhibited GUS activity, indicating that the two marker genes borne on a single plasmid are co-introduced and co-expressed at a very high frequency. This transformation procedure reproducibly yields transformants at frequencies of 5.7-6.8 x 10 ^"4 (based on the number of protoplasts electroporated) or 23.0% (based on the number of microcalli formed) counted after 6 weeks of selection. After repeated subculturing on regeneration medium, shoots are induced from 8.0% of the transformed calli. Southern hybridization confirmed the presence of both the GUS and hygromycin genes in the transformed calli and shoots. Protoplast Isolation Protoplasts are isolated from immature cotyledons of Glycine max (L.) Marr. cv. Clark 63 plants and purified as follows. Immature pods are collected from 60-80 day old plants and surface sterilized with 20% chlorox for 15 min. Cotyledons, 3-4 x 2 mm size are cut transversely into 1-2 mm thick sections and plasmolyzed for 1 hr in CPW 13M. After two washings with CPW 13M approximately 1 gm of tissue is incubated in 15 ml of enzyme solution in two 60 x 15 mm plastic petri dishes for 4-6 hr in the dark with shaking at 50 rpm. The enzyme solution consisted of 1.5% (w/v) Cellulase "Onozuka" R10, 0.2% Pectolyase Y23 dissolved in CPW 9M (pH 5.8) which is filter sterilized with a 0.2 μm pore size Nalgene filter unit. Protoplasts varied in size from 20-35 μm in diameter. The released protoplasts are filtered through 74 and 43 μm sieves and pelleted by centrifugation at 100 x g for 10 min. Protoplasts are washed twice by resuspending in KP8 medium and then purified by floating over 23% (w/v) sucrose or 13% Ficoll in CPW salts (pH 5.8) overlaid with 2.0 ml of KP8 medium and centrifuged at 80 x g for 10 min. Protoplasts are gently removed with a Pasteur pipette from the interface and washed once again with KP8 medium. Protoplast viability is determined with fluorescein diacetate (FDA; 24) test. When protoplasts are floated on sucrose (23%) there is no deleterious effect on viability but when Ficoll (13%) is used, 100% of the protoplasts burst within a few hours. Plasmid construct

The plasmid pZA300 contains a hygromycin phosphotransferase gene, known to confer hygromycin β resistance to plant cells together with the β-glucuronidase gene from pBI121 (Clonetech Laboratory, CA, USA). This chimeric gene is driven by the CaMV 35S promoter and has the nopaline synthase polyadenylation signal of pCaMVNeo. The plasmid is multiplied in the E. coli vector pTZlSR.

Electroporation

Protoplasts are resuspended at the density of 1-2 x 10 ⁶ /ml in electroporation buffer (10 mM HEPES (pH 7.2), 150 mM NaCl, 5 mM CaCl ₂ 2H ₂ O and 0.2 M nannitol). One ml of protoplast suspension is chilled briefly in ice water, then heat-shocked at 45°C for 5 min prior to addition of 20μg/ml supercoiled plasmid DNA. Subsequently, polyethylene glycol (PEG, 6000 MW; 200 μl of 28% stock) dissolved in KP8 medium containing 120 mM MgCl ₂ is mixed with the protoplast:DNA suspension. The samples are transferred to precooled, presterilized 0.4 cm plastic cuvettes fitted with aluminum foil electrodes. An electric field of 500 V/cm is applied by a single discharge of a 1000 μF capacitor that has been previously charged with an electrophoresis power supply. Four aliquots are treated with plasmid DNA and other aliquots are either electroporated without DNA or treated with plasmid DNA without electroporation to serve as controls.

Hygromycin Sensitivity of Protoplasts

Hygromycin sensitivity is determined by adding varying concentrations of hygromycin at different time intervals, starting 2 hr after culture initiation to 20 days after culture. The selection pressure is maintained for 4 weeks by replacing the liquid medium containing the antibiotic periodically (after every 7 days).

Protoplast Culture and Selection of Transformants

Following electroporation, protoplasts are cultured at 2-4 x lQr/ml density in liquid KP8 medium as follows. Liquid Culture

Protoplast density is adjusted to 2.5 x 10 ⁵ protoplasts/ml in KP8 medium containing 2% Ficoll (Type 400-DL), 40 mM MES and 2.5 ml aliquots are dispensed in each 60 x 15 mm plastic petri dish. Alternatively, protoplasts are cultured in 25 μl drops (10 drops/petri dish). During culture the osmolarity of the medium is progressively reduced by addition of 0.25 ml or 10 μl drops of a 1:1 mixture of KP8 : K8 medium to each dish at day 7 and addition of an equal volume of a 1:2 mixture of KP8 : K8 at day 10.

In liquid medium studies on the influence of initial protoplast density revealed that at densities lower than 2.5 x K^/ml, cell budding is frequently observed. Plating efficiencies of 40-50% in liquid layer, slightly lower, 38-42% in liquid microdrops are observed after 7 days in culture. In parallel, the pH value dropped to 3.9-4.1 immediately, as a result, frequency of cell division decreased dramatically and budding is observed in nearly all the dividing protoplasts. However, gradual reduction in the osmolarity stimulated development of cell colonies. Culture in agarose medium

Protoplasts are resuspended in molten KP8 medium containing 0.6 or 1.2% (w/v) Sea

Plaque agarose (at a final density of 2 x 10 ⁵ protoplasts/ml and 4 ml is plated in each 60 x 15 mm petri dish or dispensed as 25 μl droplets on die bottom of petri-dishes (10 drops/dish). After 6-8 hr each layer is cut into 4 sectors which are transferred to 90 x 15 mm petri dishes. The agarose sectors and droplets are bathed in 5.0 ml of KP8 liquid medium in each dish. During culture, the osmolarity is progressively reduced by removing the original liquid medium and replacing it with fresh KP8 : K8 medium in the ratios 2:1, 1:1, 0:1, 0:1 at day 7, 14, 21 and 28 respectively. At day 35 MSB liquid medium (0.5 mg/1 each of 2,4-D, BA, Kn and 500 mg/1 CH) is added to the cultures. After 7 days, the sectors and droplets containing micro- colonies are transferred to the surface of MSB medium solidified with 0.6% agar. When protoplasts are plated in 0.6% agarose bead sectors and droplets at 2.5 x 105/ml density protoplast division is stimulated resulting in a plating efficiency of 52-58%. The use of 1.2% agarose reduced protoplast lysis during the initial stage of culture and further increased the plating efficiency up to 55-60% . The use of agarose bead culture technique minimized pigment production and release of phenolic compounds and also offered the possibilities of removing deleterious compounds, easily stabilizing the pH and osmotic pressure, since the medium could be changed readily. Selection

The developing protoplasts are allowed to grow for 2 weeks without selection. On day 16, the protoplast derived cells, at the 2-4 cell stage, are washed with K8 medium and the selection is applied either in liquid K8 medium containing 40 μg/ml of hygromycin or the cells are resuspended in an equal volume of 1.2% Sea-plaque agarose (LMT, FMC Corp., ME, USA) gelled K8 medium containing 50 μg/ml hygromycin. The solidified agarose cultures cut into small beads are submerged in 10 ml of K8 liquid medium containing 40 μg/ml hygromycin and kept on a gyratory shaker at 40 rpm. The liquid selection medium is replaced every 7 days over a period of 6 weeks with K8 and MSB medium (MS basal salts + B ₆ organics, using different volumes of each containing 40 μg/ml hygromycin.

When the natural hygromycin-B sensitivity of isolated protoplasts and cultured cell colonies is determined, protoplasts during the initial culture stage are more sensitive to hygromycin-B than the protoplast derived multicelled colonies. Sustained development of protoplast cultures is completely inhibited by 30-40 μg/ml hygromycin, when the selection is started after 10-15 days of culture initiation. When concentrations of hygromycin-B of 20 μg/ml are used, 0.3-0.5% of the plated protoplasts grew, though slowly to the microcalli stage (0.5-1.0 mm) in 6 weeks. When selection is applied later than 20 days the efficiency of selection is greatly diminished, as 20-25% of the cells had already reached 8-16 cell stage. Therefore 40 μg/ml of hygromycin is selected as the optimum concentration for stringent selection after 15 days.

Characterization of Transgenic Clones and Calculation of Transformation Frequency Heat shock treatment given to freshly isolated protoplasts prior to electroporation increased the percentage of dividing protoplasts observed by day 14. The stimulatory effect of heat shock on protoplast division is reflected in the increased number of colonies in control cultures and of the resistant colonies which developed following exposure of protoplast derived cells to hygromycin. No colony formation occurred in the presence of 40 μg/ml hygromycin from protoplasts not treated with plasmid DNA.

On selection medium the resistant colonies continued to grow and reached a size of 100-200 μm or more after 4 weeks in culture. During selection, visible colonies developed in agarose beads and also in the surrounding liquid medium (released from the beads).

Transformed colonies are also recovered when hygromycin is applied only in the liquid medium indicating that both the selection methods can yield large number of resistant colonies. However, the advantage of embedding the protoplasts in an agarose bead over the liquid selection method is that the culture medium can be replaced without disturbing the developing colonies. The resistant colonies are easily distinguishable microscopically after 4 weeks of selection (6 weeks after electroporation) and could be scored visually after 6 weeks. The maximum number of hygromycin resistant colonies (570-685) are recovered from 1-2 x 10 ⁵ protoplasts electroporated with plasmid DNA. This gives a relative transformation frequency of 21.5-23.0% (based on the total number of colonies which grew on selective medium), and an absolute frequency of 5.7-6.8 x 10 ^"4 (expressed in terms of the number of microcolonies recovered after electroporation). These values are based on eight transformation experiments each involving 1-2 x 10 ⁶ protoplasts.

Hygromycin resistant is maintained for at least four subcultures on selective medium and is not lost after growing on hygromycin lacking medium for two subcultures (14 days each) and then replating on hygromycin-containing medium. Regeneration

Once the resistant calli grew to 1-2 mm in size they are transferred to different media for callus reformation and shoot regeneration as follows. When the microcalli grew to 1-2 mm in size, they are transferred onto MSB medium supplemented with 0.5 mg/1 each of 2,4-D, BA, Kn, 500 mg/1 CH, 3% sucrose and 0.6% agar (with or without millipore 0.45 mm pore size filters). Following regular subculturing the dark green callus pieces with nodular structures are selectively transferred to MSB medium with various combinations of auxins and cytokinins for shoot formation. For further elongation shoots are transferred in 1/2 strength MS minerals containing 1.0 gm l" ¹ KNO ₃ , 0.01 mg/1 TH with or without 0.5-1.0 mg/1 GA ₃ . For rooting 2- 3 cm long shoots are cultured in liquid or agar solidified 1/2 strength MS medium with 1 % sucrose and 0.2 mg 1 IBA or 0.5 mg/1 NAA and the cultures are initially kept in the dark for

10-12 days. Once root induction is observed these plantlets are transferred to Magenta boxes containing autoclaved vermiculite and are regularly watered with Hoagland's solution. These boxes are kept under high light and humidity conditions in the culture room. When the plants attained a height of 6-8 cm and developed two to three new trifoliate leaves they are transferred to 3:1 soil and vermiculite mixture in bigger glass bottles with loose caps to lower the humidity. Later, the lids are removed from these bottles to expose the plants to culture room conditions. Finally, the plants are transferred to soil and are kept in the greenhouse.

After 5-6 weeks of culture in KP8 medium regularly diluted with K8 medium, green microcalli (1-2 mm) could be observed in liquid layer as well as in agarose beads, which showed a survival rate of near 100% during subsequent transfer to fresh MSB medium. The calli had to be transferred to MSB medium containing phytohormones to stimulate callus growth, otherwise the calli would stop growing and turn brown. On subculture on MSB medium with 0.5 mg/1 each of 2,4-D, BA, Kn and 500 mg/1 CH with or without filter paper, the colonies grew (to 3-5 mm) within 3-4 weeks and proliferated as light green calli. When grown on millipore filter paper the calli grew faster and are more compact than the calli grown directly on culture medium. When the fast growing green calli are selectively transferred onto fresh medium of the same composition, green compact calli are formed after 3-4 subcultures. When transferred to MSB medium containing 3% sucrose and 0.5 mg/1 each of BA, Kn, ZT, 0.1 mg/1 NAA and 500 mg/1 CH and incubated under high light intensity (2,000 bu), in 2-3 weeks dark green nodular structures appeared on the surface and periphery of these calli. In one experiment an average of 15 +. 3 green shoots developed from these meristematic nodules in 21.2% of the calli after 3-4 subcultures of 14 days each. Addition of 50 mg/1 each of glutamine and asparagine and 0.2 mg/1 GA ₃ further enhanced the number of shoots to 25 +. 3 and the frequency up to 29.0%. Shoots (0.5-1.0 cm) are transferred to several different media for elongation. Elongation rate is significantly affected by growing the shoots on 1/2 strength MS minerals containing 1.0 gm/1 KNO ₃ , 0.01 mg/1 TH and 1.0 mg/1 GA ₃ . In two or three weeks approximately 60% of the cultures produced 2-3 cm long shoots.

While callus occasionally produced roots, the shoots regenerated from callus did not, without further treatment. Transfer of isolated shoots (2-3 cm) onto 1/2 strength MS minerals with 1 % sucrose, 0.2 mg/1 IBA or 0.5 mg/1 NAA resulted in the initiation of roots at the cut end of shoots in 10-12 days in the dark. The regenerated plants are gradually acclimatized to greenhouse conditions.

Protoplasts are isolated several times from the immature cotyledons obtained from greenhouse as well as field grown plants and from each experiment plating efficiencies in excess of 60% are obtained. Regenerable callus with multiple shoots and plants are recovered from most of the experiments. The calli seem to retain regeneration potential even after several

months. In general, it took us 4-5 months to regenerate plants from isolated protoplasts including rooting of the shoots.

The protoplast development reported here is similar to that reported earlier with other Glycine max genotypes where protoplast division occurred after 3-5 days with plating efficiencies of 37-41 % recorded after 7 days of culture. The reasons for the enhanced protoplast development in the present communication with another commercial genotype remains unclear, but may be caused by the combined effects of modifications in protoplast culture methods as described in this report: (i) use of 40 mM MES buffer in the culture medium to stabilize the pH in the initial stages (the pH normally dropped from 5.8 to 3.9), (ii) use of 2% Ficoll (type 400 DL) in the culture medium to prevent protoplasts from adhering to the surface of petri dishes, (iii) immobilization of the protoplasts in 1.2% agarose and initial incubation in the dark, since protoplasts placed immediately after isolation in the light bleached and many burst after 3-5 days, (iv) plating small colonies on agar solidified MSB medium with millipore filter paper to possibly adsorb phenolic substances and to facilitate gaseous exchange. Shoot Differentiation from Transgenic Clones

Protoplast derived resistant calli are usually subcultured once every 2 weeks on MSB medium containing 0.5 mg/1 each of 2,4-D, BA, Kn and 500 mg/1 CH. After 4 subcultures, calli started becoming nodular. Selective subculturing of the nodular calli on MSB medium supplemented with 0.5 mg/1 each of BA, KN, ZT, 0.1 mg/1 of NAA, 0.2 mg/1 of GA ₃ and 50 mg/1 of asparagine and glutamine, triggered the differentiation of green buds and leafy shoots after 4-6 or more subcultures. Of 440 calli derived from 34 independent transformed clones maintained on regeneration medium, 35 calli regenerated 5-20 mm long shoots (10-12 per calli) after 6-8 subcultures, giving a regeneration frequency of approximately 8.0%. Leaf samples assayed for GUS activity showed the blue color indicating that organogenesis is compatible with sustained expression of this introduced gene. GUS-Assav

GUS activity is measured by in situ staining of electroporated and non-electroporated protoplast derived cells or calli. The GUS assay buffer solution contained 100 mM sodium phosphate (pH 7.0), 0.4% X-gluc, 5 mM potassium ferrocyanide, 5 mM potassium ferrocyanide and 0.1-1.0% triton X-100. Incubation is carried out at room temperature for 12- 24 hr.

GUS activity is evident by blue staining in many cell colonies (1.0% of the total number of electroporated protoplasts) tested 14 days after electroporation even before selection for the co-introduced hygromycin-resistance gene began. On one replicate, 130 of 140 microcalli tested 14 days after hygromycin selection showed GUS activity (95% co-expression) and after 14 days of additional selection, 495 of 530 calli (0.5 to 0.8 mm diameter) tested

showed activity (94% co-expression). GUS activity is also resent in 82 of 86 callus pieces taken between 6 and 8 weeks after hygromycin selection began (95.3% co-expression). These results show that most of the selected hygromycin resistant cells also express GUS activity, thus confirming the efficient co-transformation of the linked genes in these experiments. Non- transformed cells did not show blue staining even after prolonged incubation periods of 24 hr. DNA Extraction and Southern Hybridization

DNA is prepared from twelve hygromycin resistant, GUS positive callus lines and from transgenic shoots from four of these clones by a hot phenol-lithium chloride procedure. At the LiCl precipitation step, the soluble fraction is saved and precipitated twice with ethanol. The DNA pellet is resuspended in distilled water and treated with 10 μg of RNAse to remove RNA from the sample.

DNA is digested with Hindlll or EcoRI, fractionated on 0.8% agarose and transferred to nitrocellulose filters. The filters are hybridized with ³² P labelled 2 kb HindDI or 3 kb EcoRI fragment from pZA300 containing Hpt or GUS genes. The filters are washed twice at room temperature in 2 x SSPE, 0.2% (w/v) SDS, once at 65°C and subsequently with 0.3 x SSPE, 0.1% (w/v) SDS for 1 hr at 65°C. Hybridization is visualized by exposure of the membranes to Kodak XAR-5 film at -70°C with intensifying screens.

Southern analysis of DNA from 12 independent hygromycin resistant calli and transgenic shoots from four of these cell clones showed bands of about 3 kb (EcoRi) when hybridized with the GUS gene probe and hybridizing bands of about 2 kb (Hindlll) with the Hpt gene probe. These bands are of the size expected for the Hindlll and EcoRI endonuclease treated DNA. No band is present in the lane containing DNA from untransformed soybean tissue. These results indicate that the GUS and Hpt genes have been stably integrated into the soybean genome. A detailed study of copy number and integration patterns in transgenic lines is in progress. EXAMPLE 2 Protoplast Isolation

Protoplasts were isolated from immature cotyledons of Glycine max (L.) Merr. cv. Clark 63 and purified as described in Example 1. Protoplast viability was determined by staining with fluorescein diacetate (FDA; 29). Electroporation

Purified protoplasts were resuspended at 1 x 10 ⁶ /ml density in electroporation buffer containing 10 mM HEPES (pH 7.2), 150 mM NaCl, 5 mM CaCl ₂ , 0.2 M mannitol. Supercoiled plasmid DNA (20 μg/ml) in TE buffer was then added. No carrier DNA was used in these experiments. Electroporation was performed at 500 V/cm, 1000 μF using a single pulse as described in Example 1.

Plasmid constructs

The BAMHI - Kpnl T _R fragment from pSaK5 (pSa4 vector containing Kpnl fragment 5 from pTil5955, S.K. Farrand, unpublished data) was cloned between the unique BAMHI - Kpnl sites of pUCD2001, resulting in the T-DNA vector pMAS2. The insert includes T _R , R _c and part of the T _L - region of pTil5955 (Baker, R.F. et al., (1983) Plant Mol. Biol. 2:335-350). The T _R -region contains ORFs 24, 25 and 26 corresponding to transcripts 2', 1' and 0', which are necessary for the biosynthesis of the mannityl opines (Ellis, J.G. et al., (1984) Mol. Gen. Genet. 195:466-473). An NPTII gene fused to the 24S promoter of the CaMV (Fromm, M.E. (1986) Nature 319:791-793) was cloned into the Xbal site located in T _R ORF 21 of pMAS2 (Baker, et al. supra). Plasmid pUSD2001 also contains an Xbal site located in the pTAR par region (Galli, D.R. et al., (1985) Plasmid 14:171-175). During the cloning, the 5.9 kb fragment between the Xbal sites in ORF 21 and the par locus of pMAS2 was deleted. In plasmid screenings, a single recombinant clone, pMAS4 was isolated which contains the T _R right border, ORFs 24, 25 and 26, and the NPTII cassette. The 3' 456 bps of ORF 21 corresponding to gene 3' remain in pMAS4. However the 5' structural and regulatory sequence of this gene have been deleted. This plasmid lacks about 200 bps of the pTAR par locus and the 5.7 kb BAMRl - Xbal fragment of the T-region including the R _R left border, T and the T _L segment of Kpnl fragment 5. The deletion has no detectable effect on the stability of MAS4 in Agrobacterium or E. coli hosts. Co-transformation

Co-transformation experiments were performed essentially as described above using pCaMVNeo (4.4 kb, 13) and pMAS2 plasmids. DNA concentrations, at ratios of 11 or 1:5 of pMAS2 and pCaMVNeo were mixed in 100 μl TE buffer and added to 1 x 10° protoplasts in 1 ml of electroporation buffer. Protoplast culture and selection of kanamycin-resistant colonies

Electroporated protoplasts were resuspended in Kp8 medium at densities from 2-4 x 10 ⁵ /ml and were allowed to grow for two weeks without selection. On day 15 protoplast-derived cells, at the 2-4 cell stage, were washed with K8 medium and resuspended in an equal volume of 1.2% agarose medium (5 ml) containing 50 μg/ml kanamycin sulfate (Sigma). The agarose cultures were cut into slices and allowed to grow for 6 weeks under the antibiotic selection in K8 and MSB medium as described in Example 1. Some of the electroporated protoplast-derived colonies were maintained in medium as described above without exposure to kanamycin to ensure that these samples were capable of colony formation.

The absolute transformation frequency (ATF) was calculated as the number of resistant colonies produced after 8 weeks divided by the initial number of protoplasts plated after electroporation. The relative transformation frequency (RTF) is described as the ratio between

the number of resistant colonies in selected cultures and the developing colonies in unselected cultures.

Plant regneration

After 8-10 weeks of selection, individual kanamycin resistant callus pieces, 1-2 mm in size, were transferred to different media as described in Example 1 for callus formation and shoot and root induction. The regenerated plants were transferred to pots containing vermiculite and soil mixture.

Neomycin phosphotransferase II (MPTII. assay

The NPTII enzymatic activity was qualitatively detected in callus and leaves from individual transformed clones by a dot blot procedure. Tissue (100 mg fresh weight) extracts were made in 50 μl of extraction buffer. (See McDonnell, R.E. et al., (1987) Plant Mol. Biol.

Rep. 5:380-386) After incubation of the cell extract with reaction buffer twenty microliter of reaction mixture was blotted onto Whatmann ™ p81 (cellulose phosphate) paper, the paper was washed, dried and exposed to X-ray film for 24 hr at -70°C. Opine assay

For the detection of mannityl opines, 30-50 mg fresh weight of kanamycin-resistant callus, leaf or root tissue from individual plants was homogenized in 50 μl of 70% ethanol containing 5 μl of electrophoresis running buffer (formic acid:acetic acid: water, 3:6:91, v/v/v, pH 1.9). Tissue extract was spotted onto Whatman No. 3 filter paper and opine were separated and visualized as described by Savka et al., (1990) Phytopathology 80:503-508.

Southern hybridization

DNA was extracted from callus or leaf tissue essentially as described by Dellaporta et al., (1983) Plant Mol. Biol. Rep. 1:19-21. Ten μg of DNA digested with Xbal was electrophoresed in a 1.0% w/v agarose gel. The DNA was blotted onto a nitrocellulose filter and hybridized with the ³² P-labelled 1.7 kb Xbal fragment containing the NPΗI gene from pMAS4. The filter was washed twice with a solution of 2X SSPE and 0.5% w/v SDS for 15 min at room temperature followed by 15 min at 68°C. Subsequently, the filter was washed with 0.2X SSPE and 0.2% SDS for 15 min at 68°C with agitation, dried and exposed to X-ray film using an intensifying screen for 24 to 72 hr at -70°C. Results and Discussion

Selection of transformants and calculation of transformation efficiency Protoplasts prepared from immature seed were electroporated with pMAS4 plasmid carrying kanamycin resistance and opine synthesis genes and the transformants were selected on kanamycin medium. No resistant colonies were recovered if kanamycin selection (at 50 μg/ml) was applied directly at the time of culture or to one-week-old protoplast derived colonies. Only a few protoplasts (< 1%) divided when 25 μg/ml of kanamycin was added one week after

culture initiation, but no development of such cells was seen beyond the eight-celled stage. Growth of 14 d old protoplast-derived cells was inhibited by 50 μg/ml of kanamycin, but exposure to kanamycin after three weeks did not stop colony growth even at 100 μg/ml. When selection was started after two weeks, 2-4 celled colonies had formed, which seems to be the optimal growth stage of the cells for beginning selection. Therefore, transformed cells were selected by incorporation of 50 μg/ml of kanamycin into the liquid medium bathing the garose sectors 14 d after culture initiation.

The viability and division frequency of protoplasts subjected to electroporation (1000 μF, 500 V/cm, one pulse) was generally lower (55-60%) than that of untreated control protoplasts (10). Electroporated protoplasts plated in agarose readily formed microcolonies within 8-10 days after transfer of the agarose slices to liquid medium containing 50 μg/ml kanamycin. On selection medium, the resistant colonies continued to proliferate and were easily distinguishable visually from the non-transformed cells after 6 weeks. The maxiumum number of kanamycin resistant colonies recovered varied from 370 to 460 from 1 x 10^ protoplasts/ml in four experiments with each two individual treatments. This gives the relative and absolute transformation frequencies of 12.4-14.7% (based on the total number of colonies which grew on the selection versus the non-selective medium) and 3.7-4.6 x 10 ^"4 , respectively. Protoplasts electroporated in the absence of plasmid DNA failed to grow on kanamycin containing medium. Approximately 95% of the resistant colonies were transferred to MSB medium grew and increased in size to more than 2-5 mm within one month. Kanamycin selection (at 50 μg/ml) was maintained for at least four subcultures and resistance was not lost even after growing the callus on medium lacking kanamycin for two subcultures (two weeks each) and then replating on kanamycin containing medium. Analysis of enzyme activity in the transformants In one experiment (using 1 x 10 ⁶ protoplasts/ml) with 32 kanamycin resistant cell lines randomly chosen after 12 weeks of selection, 27 (84%) were NPΗI positive and of these 24 (75%) showed presence of opines. Most of the opine-positive lines produce agropine, mannopine, mannopinic acid and agropinic acid. In three different experiments with a total of 54 colonies, 45 were NPTII positive and 40 were opine positive. Thus not all of the resistant colonies selected on kanamycin showed the presence of

NPTII and the opine accumulation. It is possible that after reaching a certain minimum size non-transformed colonies in soybean can exhibit tolerance to kanamycin at the concentration used in the selection medium. Such size related differential antibiotic resistance has been observed with other selection systems also. However, in this system the recovery of escapes can be avoided by applying a more stringent selection using kanamycin at 75 μg/ml. In one experiement, colonies were selected on 75 g/ml of kanamycin and sixteen randomly chosen

growing colonies were tested and all of them contained both the NPTII activity and accumula_ed opines, giving a co-expression frequency of 100%. Plant Regeneration

All of the isolates (clones) obtained by selection with 50 ug/ml kanamycin which were positive for NPTII activity and opine accumulation were multiplied. Fast growing calli were repeatedly subcultured on regeneration medium as described in Example 1. Approximately 10% of the dark green nodular calli of 10-15 mm size produced an average of 8-10 shoots (each of 10-20 mm size), of which approximately 40% could be elongated. These elongated shoots were rooted to produce complete plantlets, which were later transferred to a vermiculite and soil mixture. The plants obtained in this study have been placed under the appropriate conditions for flowering and seed production. Southern Hybridization Analysis

To obtain molecular evidence for stable transformation, total genomic DNA was isolated from callus and plants regenerated from different cell clones which were NPΗI and opine positive and from control soybean plants. Genomic DNA was treated with Xbal which produces a 1.7 kb NPTII gene gragment from the plasmid pMAS4. The Southern blot containing the transferred genomic DNA was hydridized with radioactive probe prepared from 1.7 kb NPTII coding region.

Indeed, all of the 32 independent transgenic callus clones and 11 independent transgenic plants produced from these clones (NPTII and opine positive) analysed so far showed hybridization to the expected 1.7 kb gragment from Xbal - digested pMAS4 DNA, while DNA from untransformed callus and leaves did not contain sequences that hybridized with the probe. When the uncut DNA from selected cell lines was probed with the NPTII gene fragment, hybridization was only seen in high molecular weight DNA, indicating that the plasmid had integrated into the genome. The Southern blot experiments on transgenic calli and plants confirmed the stable integration of the transformed DNA. The DNA patterns in the calli and the regenerated plants were similar, indicating that no rearrangements have taken place during the culture and induction of organogenesis in the calli. The evidence demonstrating the presence of the transforming DNA, resistance to kanamycin and the accumulation of opines were complemented by enzymatic proof that the gene is functional, since the specific phosphorylation of the kanamycin from the tissue extracts prepared from the transgenic plants. The plants taken for the enzyme assay were the same as those used for the Southern analysis. All the transgenic plants (seven to date) analysed for NPΗI activity and opine accumulation tested positive. The detection of the NPTII activity and opine accumulation in all of the resistant clones indicated that transfer of both marker genes had occurred into the soybean genome. Molecular analysis confirmed the integration of both genes into genomic DNA <

the copy number estimated to be 1-3 copies per haploid genome. Single Plant Assay

Before being transferred to soil, elaves and roots form individual transgenic plants which showed the integration of SPTII and opine genes into their genome were assayed for corresponding enzyme activity and opine accumulation. Seven to thirteen leaves from seven individual plants at different developmental stages (total of 72 trifoliate leaves, each of 20-50 mg fresh weight) were found to contain opines in their extracts as did extracts of roots of the same plants. Likewise the extracts from all the 72 leaves and roots tested were positive for NPΗI activity when dot blot assay was carried out. Co-transformation with the NPTII and Opine svstensis genes on separate plasmits

In co-transformation experiments two donor plasmids were used: pCaMVNeo (Fromm, M.E. et al., supra) which contains the NPTII gene flanked by the 35S cauliflower mosaic virus promoter and the nos 3' terminator and pMAS2 which contains mannityl opine synthesis gene, using pCaMVNeo and pMAS2 plasmid DNA ratio at a ratio of 1:1 (10 g/ml of each), the number of kanamycin resistant colonies ranged from 185 to 200 from 1 x 10 ⁶ protoplasts. In three different experiments, a total of 58 kanamycin resistant lines were analysed for NPΗI enzyme activity and 49 colonies tested positive (in 3 replicates), indicating that 85% of the resistant colonies expressed the introduced NPTII gene. When extracts from all the 58 colonies were assayed for opine accumulation, 38 contained the opines, which showed that the expression of unlinked genes occurred in about 65% of the kanamycin selected colonies. To determine the effect of DNA concentration on transformation efficiency the concentration of selectable marker pCaMVNeo was increased from lOμg to 50 μg/ml in combination with lOμg/ml pMAS2 (in 50 μl TE buffer) this resulted in a slightly increased number of transformed colonies (185 to 245 recovered from 10 ⁶ protoplasts upon selection on kanamycin medium). Simultaneously, the number of NPTII expressing colonies also increased (in one assay 18 of 20 colonies were NPΗI positive showing 90% expression) but the fraction of opine expressing colonies did not change, remain nearly 65% of the selected cell clones. Likewise when the concentration of pMAS2 to pCaMVNeo from 10 μg to 50 μg/ml, the number of opine expressing colonies also increase. These results indicate that linkage of the selectable and non-selectable marker on single

NA molecule leads to a higher transformation frequency (85%) in comparison to co- transformation of two unlinked genes on separate plasmids (65% efficiency). The co- transformation frequency reported here is higher than the frequencies of 25-59% described with other genes in Arabidopsis, maize and rice using PEG mediated transformation. In Example 1 the expression of linked chimeric genes (hygromycin phosphotransferase and jS-glucuronidase from the 7.7 kb plasmid pZA300, using 20 μg/ml) at the rate of more than

93%, by selecting the transformants with 40 g/ml of hygromycin in soybean is reported. However, using a larger plasmid (1" . kb) with two genes linked together (neomycin phosphotransferase and amnnityl opine synthesis) the expression was somewhat lower (85%). Larger plasmids may possibly be subject to more fragmentation during electroporation or in the cell. Hence, it could be assumed that plasmid size per se and length of the gene of interest within the vector is critical for stabilizing DNA molecules during the process of transformation with small plasmids being more stable than the large ones. Conclusions

Kanamycin resistant plants were regenerated from Glycine max protoplasts electroporated with a plasmid vector containing linked chimeric genes with 85% co-expression. However, if the genes are on separate plasmids the efficiency of co-expression is reduced to about 65%. In selection of microcalli, we found that 75 μg/ml kanamycin should be used to prevent escapes. Both NPTII and all the genes from mannityl opine biosynthetic region of pMAS4 plasmids were present in different organs of transgenic plants as shown by enzyme and opine assays and by Southern hybridization indicating that nonchimerial plants are produced. EXAMPLE 3 Plant Material

Plants of fourteen cultivated Glycine max genotypes (Table 1) were grown in the field or under greenhouse conditions (16 hr supplemented light per day, 26+2°C). Protoplast Isolation and Purification

Immature pods from 60-80 day old plants were collected and surface sterilized with 1 % sodium hypochlroite (20% clorox) for 15 min. Young immature cotyledons of 4 x 2 size were dissected from the pods. After removing the seed coat, excised cotyledons were cut transversely into 0.5-1.0 mm thick sections and plasmolyzed for 1 hr in CPW 13M (Power and Davey, 1980). After two washings with CPW 13M, approximately 1 gm of tissue was incubated in 15 ml of enzyme solution containing 1.5% (w/v) Cellulase 'Onozuka R10', 0.2% (w/v) Pectolyase Y-23 and 9% w/v) mannitol (pH 5.8) on a gyrotory shaker (50 rpm) in the dark for 4-6 hr. The procedure for protoplast purification ws the same as described in Example 1. Protoplast Culture

Purified protoplasts were resuspended in different culture media (see Table 2) with or without LMP (Low Melting Point, FMC Corporation, Michigan, MI) agarose at a final density of 2 x 10 ⁵ protoplasts/ml. The agarose medium was allowed to solidify for 6-8 hr at room temperature and was later cut into four sectors. These sectors were then transferred to a 100 x 15 mm petri dish containing 10 ml of the liquid medium.

During culture, using either of the two techniques (liquid or agarose bead), the

osmolarity of the medium was progressively reduced by adding fresh medium or diluting it with K8 medium (Kao, 1977) as described in Example 1. For the first 7-9 days, petri dishes (in loosely sealed plastic boxes) were kept in the dark at 26+2°C. The plating efficiency (number of dividing protoplasts expressed as a percentage of the initial protoplast population) was determined 7 days after culture. Later, cultures at different developmental stages were gradually transferred to high intensity light conditions. Plant Regeneration

Once small colonies (1-2 mm in size) could be observed visually they were picked up carefully from the beads or were released from the agarose matrix by applying pressure to the beads with a spatula. Colonies released into the surrounding liquid medium or picked up from the beads were further subcultured onto MSB medium containing MS (Murashige and Skoog, 1962) salts and B5 (Gamborg et al., 1968) organics supplemented with 0.5 mg l ^"1 CH, 3% sucrose and 0.6% agar or 0.1% gelrite. These colonies were subcultured 3-4 times at 14 day intervals. When these colonies appeared green and compact they were transferred onto regeneration medium consisting of MSB medium with 3% sucrose, 0.5 mg l ^"1 each of BA, KN, ZT, 0.1 mg l ^"1 NAA, 500 mg l ^"1 CH, 50 mg l ^"1 each of asparagine and glutamine and 0.2 mg l ^"1 GA ₃ Besides MSB regeneration medium (used in Example 1), four other different media i.e. TH (MS standard medium containing 0.3 mg l ^"1 IBA and 0.03 mg l ^"1 thidiazuron), OR and MSR (Barwale and Widholm, 1990); B5 (Gamborg et al., 1968) with different sugars and phytohormone combinations were also tried. Cultures were kept under high intensity light until dark nodular structures appeared on the periphery of calli. Later, these nodular structures were dissected out and gently washed with liquid medium of the same composition to remove dead tissues and phenolic compounds. These selected structures with green bud primordia were transferred onto fresh MSB medium of the same composition for 3-4 subcultures before transferring onto shoot elongation medium as reported in Example 1. Light intensity was increased gradually throughout the regeneration procedure from 10 μE m ^"2 .s ^_1 to 25-30 μE m ^{" 2} .s ^_1 . Elongated shoots of 2-3 cm length were excised and transferred to half strength MS minerals with 1 % sucrose and 0.5 mg l ^"1 NAA, for root induction. The regenerated plants, so produced, were transplanted to pots containing a vermiculite:soil mixture or soil and later were transferred to the greenhouse.

Protoplast Culture and Plating Efficiency

Protoplasts were readily isolated from all fourteen soybean genotypes included in this study using 1.5% Cellulase and 0.2% Pectolyase within 4-6 hr of incubation. Usually about 5 x 10 ⁶ protoplasts were released from one gm fresh weight of immature cotyledon tissue. However, the yield varied with the genotype and ranged from 1-8 x 10 ⁶ protoplasts per gm fresh weight of tissue (Table 1).

Viability of isolated protoplasts varied among experiments and genotypes, but was generally over 70% as indicated by FDA staining and protoplast morphology, except in Burlison and XP 3015 where the viability was 60-70% (Table 1).

In initial experiments protoplasts of all the cultivars were cultured (at a density of 2 x 10 ⁵ ml) in KP8 liquid medium alone or with 1.2% LMP agarose as was used for Clark 63. First cell divisions were observed with 48 hr when the protoplasts were embedded in agarose and in 96 hr when cultured in liquid medium. Protoplast divisions counted after 7 days showed that, not all of the genotypes had high plating efficiencies especially when cultured in liquid medium, as summarized in Table 1. However, protoplasts of 8 from the 14 genotypes (cultured in KP8 medium) showed 40-63 % plating efficiency, indicating that approximately 60% of the genotypes tested had the same sugar requirement for initial protoplast divisions. In the remaining six genotypes very few initial divisions (colony up to 2-cell stage; less than 100 divisions per 2 x 10 ⁵ protoplasts) were observed during the first 7 days of culture. After 7-10 days, the protoplasts started to die leading to complete death in about two weeks. For the protoplasts of the six genotypes which did not respond on KP8, we tried different media, i.e. B5, MS and KP8 with varying carbon sources (D1-D9 medium, see Table 2). On these media protoplasts of the various genotypes gave plating efficiencies in the range of 38-60% after 7 days of culture (Table 1). These results indicate that each genotype has its specific sugar and salts requirement in the early stages of development, as is evident by growing protoplasts on different media. After 7 days of culture, cells on compatable medium had undergone 2-3 cycles of division. Genotypic differences were observed in plating efficiencies with the highest rate (63%) for A-2396 and Jack while X-3337 exhibited the lowest rate (38%). When the protoplasts were culture in liquid medium, most of the genotypes showed low plating efficiencies and accumulated brown pigments in the cell wall and eventually the culture medium turned brown as well. Embedding the protoplasts in agarose beads (0.6%) did not improve viability and the division frequency was similar to that in liquid culture medium. However, embedding protoplasts in 1.2% agarose produced better results, as more than 80% of the initial protoplasts remained intact after plating, which proved to be advantageous for inducing initial divisions as well as for fast growth of the protoplast derived colonies. When the agarose bead method is used it is also easier to change the medium without disturbing colonies as reported in Example 1. Callus Formation

The cultures were diluted with the D1-D9 medium, optimal for each genotype and K8 medium at a 1:1 ratio on day 8 and 14, 1:2 on days 21, and 28 and then K8 medium alone was added after 5 weeks as reported in Example 1. Upon dilution, cells of the genotype Jack divided faster than other genotypes and grew to 64 or more cells in 3 weeks of culture.

Colonies that formed in agarose from all genotypes were organized and compact, whereas those formed in liquid medium were loose, unorganized and slow growing. After 5-6 weeks of culture, green microcalli (1-2 mm in size) could be observed. Fast growing green colonies were picked and subcultured onto MSB medium with 0.5 mg l ^"1 each of 2,4-D, BA and KN, 500 mg l ^"1 CH (after every 14 days each). After 3-4 regular subcultures of 14 days each these calli grew further (5-10 mm in size) and became bright-green, nodular and compact. Friable tissues continued to proliferate as friable callus and never produced shoots. During this culture period, calli that had reached a size of 8-10 mm were transferred onto solid shoot regeneration medium. No morphogenesis was observed if the callus pieces were left on the same medium for more than 14 days before subculturing. These calli began to turn brown at the periphery and their color changed to yellowish brown. Regeneration of Plantlets and Their Growth

Nodular, bright green and compact calli obtained from all the fourteen genotypes were transferred onto MSB regeneration medium (Table 3) and were maintained under higher intensity light (25-30 μE m ^"2 .s ^-1 ). After 4-8 regular subcultures, 14 days each on the same composition fresh medium, only six genotypes i.e. A-2396, Chamberlain, Heilong-26, Jack, Resnick and XP-3015 regenerated shoots (4-10 mm in length). The genotype Jack had the highest shoot regeneration frequency of 27% (as 52 out of 192 callus pieces produced 8-10 shoots/calli, with 2-3 leaves). In addition to the six genotypes, calli from A-5403, Tiefeng and X-3096 also produced dark green nodular structures with leaf primordia in a few cultures, and s ometimes water soaked translucent leaves, but these structures did not grow further even after repeated subculturing. Four other basal media i.e. OR, MSR, TH and B5 were compared to MSB regeneration media. As the data in Table 4 show, none of these media induced regenerative structures. B5 media was actually inhibitory and the callus pieces growing on it completely turned friable and yellow-brown. However, on MSRA and MSRB medium calli became compact and produced long roots. Although the callus tissue growth on MSR medium was lower than that of the tissues growing on TH.

Isolated shoots with 2-3 leaves from the regenerable genotypes except Heilong-26, which were lost due to contamination were transferred onto shoot elongation medium containing half strength MS medium salts + 1 g l ^"1 additional KNO ₃ + 0.01 mg l ^"1 thidiazuron + 1.0 mg l ^"1 GA ₃ as reported in Example 1. drwe 2 weeks, approximately 25-45% of the shoots elongated with 3-5 true leaves. When isolated shoots, 2-3 cm in height, were transferred onto half strength MS minerals with 1 % sucrose and 0.5 mg l ^"1 NAA and grown with rooting medium shaded in the dark for 2-3 weeks, roots were observed in approximately 65-83% of the cases at the cut ends. Large number of plantlets can be produced from regenerated shoots of different genotypes. However, because of amount of labor, greenhouse space and time needed

to obtain large number of plants, a total of sixty-three plantlets have been produced so far and thirty-five plants have been successfully transferred to pots containing vermiculite and soil or only soil in the greenhouse conditions. To date, eight Jack, five Chamberlain and three A-2396 plants have set seeds. The number of seeds obtained per plant varied between 40-65 and other plants are at different stages of development.

Protoplasts have been isolated several times using immature cotyledons of fourteen soybean genotypes grown in greenhouse and field grown plants and from protoplast derived Jack plants growing in the greenhouse. No variation in plating efficiency has been observed. In all the five genotypes that have produced shoots, protoplast-derived calli seem to retain regeneration potential, as even after several months of subculture these calli are still producing multiple shoots. Conclusion

The results obtained in this study demonstrate that the plant regeneration system reported in Example 1 using the soybean genotype Clark 63 from protoplasts using immature cotyledons, can now be applied to several other commercial soybean genotypes (only after minor modifications) to recover fertile plants.

Table 1. Yield, viability and plating efficiency of protoplasts isolated from immature cotyledons of 14 different soybean genotypes

Protoplasts of all the fourteen genotypes were resuspended at 1-5 x 10 ⁵ protoplasts per ml and cultured in D1-D9 medium using liquid medium or embedding the protoplasts in 1.2% L P agarose. Plating efficiency was determined as the % o dividing protoplasts after 7 days of culture.

Table 2. Sugar composition of different media used for protoplast culture of soybean genotypes

Medium Basal Sugar composition (g 1 ) designation media Glucose Sucrose annitol Sorbitol

KP8 = Kao medium ( 1977); MS = MS medium without hormones (Murashige and Skoog, 1962); B5 = Gamborg medium (Gamborg et al., 1968). Each medium was supplemented with 2, 4-D; 0.2 mg1 \ NAA, 1.0 mg 1 \ ZT 0.5 mg 1 \ 2% Ficoll and 40mM MES buffer(pH 5.7) and was filter sterilized using 0.22 μm pore size cellulose acetate membrane filter.