Winter Survival of Transgenic Alfalfa Overexpressing Superoxide Dismutase

Abstract

To test the hypothesis that enhanced tolerance of oxidative stress would improve winter survival, two clones of alfalfa (Medicago sativa) were transformed with a Mn-superoxide dismutase (Mn-SOD) targeted to the mitochondria or to the chloroplast. Although Mn-SOD activity increased in most primary transgenic plants, both cytosolic and chloroplastic forms of Cu/Zn-SOD had lower activity in the chloroplast SOD transgenic plants than in the nontransgenic plants. In a field trial at Elora, Ontario, Canada, the survival and yield of 33 primary transgenic and control plants were compared. After one winter most transgenic plants had higher survival rates than control plants, with some at 100%. Similarly, some independent transgenic plants had twice the herbage yield of the control plants. Prescreening the transgenic plants for SOD activity, vigor, or freezing tolerance in the greenhouse was not effective in identifying individual transgenic plants with improved field performance. Freezing injury to leaf blades and fibrous roots, measured by electrolyte leakage from greenhouse-grown acclimated plants, indicated that the most tolerant were only 1°C more freezing-tolerant than alfalfa clone N4. There were no differences among transgenic and control plants for tetrazolium staining of field-grown plants at any freezing temperature. Therefore, although many of the transgenic plants had higher winter survival rates and herbage yield, there was no apparent difference in primary freezing injury, and therefore, the trait is not associated with a change in the primary site of freezing injury.

Winter hardiness is a complex trait that involves tolerances to freezing, desiccation, ice encasement (severe anoxia), flooding (milder anoxia), and disease. The combinations and severity of these stresses that crops must tolerate vary with environment and year. There are distinct differences among the environmental stresses that occur during winter, but oxidative stress may be a common element. Freezing, anoxia, and desiccation stresses have been linked with oxidative stress in three types of correlative physiological and biochemical studies. First, degenerative reactions associated with anoxia (Hetherington et al., 1987, 1988; Monk et al., 1989), desiccation (Senaratna et al., 1987), and freezing (Kendall and McKersie, 1989) are similar to the reactions caused by the herbicide paraquat (Chia et al., 1982; Bowler et al., 1991) and the pollutant ozone (Van Camp et al., 1994). Second, microsomal membranes from acclimated plants are more tolerant of in vitro, free-radical treatment than those from nonacclimated plants (Kendall and McKersie, 1989). Third, as plants acclimate to low temperatures, they acquire coincidentally increased tolerance to freezing stress, ice-encasement stress, and free-radical-generating herbicides (Bridger et al., 1994).

We hypothesized that enhancing a plant's tolerance of oxidative stress would improve its ability to survive the combination of stresses associated with winter. The mechanisms to detoxify oxygen radicals are varied and the complex interactions among the antioxidants in different subcellular compartments, cells, and tissues are only now being elucidated (Bowler et al., 1992; Herouart et al., 1993; Scandalios, 1993; Foyer et al., 1994; Allen, 1995). SOD is an essential component of these defense mechanisms because it dismutates two superoxide radicals to produce hydrogen peroxide and oxygen. Previously, a Mn-SOD cDNA from tobacoo was introduced into alfalfa (Medicago sativa L.); one of the primary transformants and its F₁ transgenic progeny showed increased survival and vigor after exposure to sublethal freezing stress in the laboratory (McKersie et al., 1993). Two of the transgenic plants also had relatively increased tolerance of water deprivation and four had increased vigor and winter survival in the field (McKersie et al., 1996). The previous studies used only a few transgenic plants because of limitations of space and facilities, and only a single alfalfa clone, called RA3, was used because it was one of the few genotypes at that time that could be transformed, although it had poor agronomic performance and was not well adapted to winter conditions.

In this study we expanded our original observations by transforming two elite alfalfa plants that are adapted to our field environment, and we examined many more independent transgenic plants in both laboratory and field evaluations. The results confirm that overexpression of Mn-SOD in transgenic alfalfa plants often improves the winter survival and subsequent herbage yield of this crop, but in some independent transgenic plants, winter survival and subsequent yield actually lessened. The transformation event had a greater influence on winter survival and shoot dry-matter yield than the subcellular site of SOD targeting. However, the improvement in winter survival of the whole plant was not related to a change in the primary freezing tolerance of the cells in the taproot or crown of transgenic alfalfa.

MATERIALS AND METHODS

Plant Transformation

Two different clones of alfalfa (Medicago sativa), designated N4 and S4, were selected from the University of Guelph plant breeding program to be transformed based on their field performance (Bowley et al., 1993). Petiole explants of alfalfa were cocultivated with an overnight culture of Agrobacterium tumefaciens C58C1 Rif pMP90 containing the binary vectors pMitSOD or pChlSOD described previously (Bowler et al., 1991). The explants were cocultivated for 3 d in the dark on SH induction medium (Shetty and McKersie, 1993) containing 288 mg L⁻¹ Pro, 53 mg L⁻¹ thioproline, 4.35 g L⁻¹ K₂SO₄, and 100 μm acetosyringinone. The explants were washed in half-strength Murashige-Skoog medium (Murashige and Skoog, 1962) and plated on the same SH induction medium without acetosyringinone but with 500 mg L⁻¹ claforan and 50 mg L⁻¹ kanamycin. After several weeks somatic embryos were transferred to BOi2Y development medium (Bingham et al., 1975) containing no growth regulators, no antibiotics, and 50 g L⁻¹ Suc. Somatic embryos were subsequently germinated on half-strength Murashige-Skoog medium. Rooted seedlings were transplanted into pots containing Turface (Plant Products, Mississauga, Ontario, Canada) in a greenhouse at approximately 23°C/18°C (day/night) and a minimum 16-h photoperiod.

PCR Screening

Prior to transfer to the greenhouse, the putatively transgenic plants were screened for the presence of the nos-nptII transgene using PCR. DNA was extracted with 400 μL of homogenizing buffer (250 mm NaCl, 25 mm EDTA, 0.5% SDS, 200 mm Tris-HCl, pH 7.4). The supernatant of a 13,000g centrifugation was mixed with 300 μL of isopropanol. DNA was collected at the interface, washed, and resuspended in water. The quality and concentration of the DNA was confirmed using a 0.8% agarose gel with ethidium bromide staining. For the PCR reaction 25 ng of DNA was combined with 1.5 μL of 15 mm MgCl₂, 1 unit of Taq polymerase, 2.5 μL of 10× buffer, 2.5 μL of deoxyribonucleotide triphosphate, and 2 μL of each primer made to a final volume of 25 μL with water. The primers used were 5′-AGCTGTGCTCGACGTTGTCAG-3′ and 5′-GGTGGGCGAAGAACTCCAGCA-3′. The PCR program included 5 min at 94°C, then 25 cycles of 94° for 15 s, 65° for 30 s, and 72° for 60 s, followed by 5 min at 72°C and holding at 4°C. PCR products were visualized on a 0.8% agarose gel with ethidium bromide.

Southern-Blot Hybridization

We treated the purified DNA from alfalfa with the restriction enzymes EcoRV and EcoRI at a 5- to 10-fold excess. All samples were separated using a 0.8% agarose gel. After electrophoresis the gel was blotted overnight onto a positively charged nylon membrane as described by Ausubel et al. (1991). After blotting the membranes were UV cross-linked. Subsequent Southern-blot analysis was based on the digoxygenin chemiluminescent system (Boehringer Mannheim) (van Miltenburg et al., 1995). We synthesized the digoxygenin-labeled DNA hybridization probes using the Expand enzyme to enzymatically label PCR products with digoxygenin-dUTP (PCR digoxygenin probe synthesis kit, Boehringer Mannheim) as described by the manufacturer. Probes were synthesized for both the kanamycin and the mitochondrial SOD genes, using 200 pg of purified plasmid DNA as the template. The PCR primers for synthesis of the kanamycin probe were 1 μm each of 5′-AGCTGTGCTCGACGTTGTCAC-3′ and 5′-GGTGGGCGAAGAACTCCAGCA-3′. The annealing temperature was 65°C and the product size was 732 bp. The primers for the SOD gene were 5′-GAGCAGACGGACCTTAGC-3′ and 5′-AGAAACCAAAGGGTCCTG-3′, with a 55°C annealing temperature and a 511-bp product.

SOD Activity

We extracted SOD from two to three fully expanded leaf blades (or other tissue as indicated) from a vegetative stage shoot. The sample was frozen in liquid nitrogen, ground, and resuspended in 150 μL of 50 mm KH₂PO₄, pH 7.8. We centrifuged the homogenate at 13,000g for 15 min and determined the protein content of the supernatant (Bradford, 1976). A constant volume (20 μL) was applied to a 13% polyacrylamide gel with a 4% stacking gel (McKersie et al., 1993). One lane of each gel contained 0.5 unit of bovine Cu/Zn-SOD (Sigma) as an internal standard. The gel was stained with nitroblue tetrazolium and riboflavin (Sigma) at 4°C and then developed on a light box for 20 min. The areas of SOD activity were clear against a blue background.

An image of the gel was captured using a CCD (charge-coupled device) video camera and Northern Exposure Software (Empix Imaging, Mississauga, Ontario, Canada). We used Excel (Microsoft, Redmond, WA) to calculate the area under each SOD isozyme peak and expressed the data as a percentage of the total activity by calculating the area of isozyme peak per total area of all SOD peaks times 100%. Alternatively, we expressed the data as units of SOD activity per gram of protein, calculated as the area of individual peak per area of internal standard times the concentration of the internal standard.

Freezing Tolerance

The transgenic plants were propagated by node cuttings and rooted in Turface growth medium in 4- × 4- × 11-cm four-cell root trainers (Plant Products). The plants were defoliated and grown to a height of about 10 cm (approximately 2 weeks after defoliation). The plants were then transferred to a 2°C growth chamber with a 12-h photoperiod, at a light intensity of 200 μmol m⁻² s⁻¹ (PAR) and acclimated for 2 or 4 weeks at these conditions. Subsequently, the 4-week-acclimated plants were frozen at −2°C overnight in the dark and then acclimated at 2°C for another week in the growth conditions described above. We sampled the leaf blades and fibrous roots at each stage of acclimation. The tissues were placed in glass tubes in a precooled aluminum block, frozen at −1°C for 1 h, inoculated with an ice chip, and cooled at a rate of −2°C h⁻¹. One sample (tube) was removed at −6°C, −7°C, −8°C, and −9°C. We thawed the sample overnight at 2°C, added 10 mL of water to each tube, and measured the conductivity after 1 h and after autoclaving at 120°C for 15 min to calculate the percentage of leakage. We repeated each experiment three times on three separate days and analyzed the data statistically as a split-plot arrangement with stages of acclimation as the main plot.

Field Trials

The 1996 field trial was conducted at the Elora Research Station (Elora, Ontario) following protocols authorized by the Plant Products Division, Agriculture and Agri-Food government agency (Ottawa, Ontario, Canada) (tests 96-UOG2-ALF03-ON0-1-01 and 96-UOG2-ALF04-ON01-01, approved April 9, 1996). The plots were established in the spring of 1996 by transplanting rooted cuttings of each transgenic and control genotype. The soil at this location is a clayed brunisolic gray-brown luvisol-London. Fertilizer (P and K) was applied after each harvest according to the results of the soil test analysis. We arranged the test in a randomized complete block design with 15 cuttings of each control (nontransgenic) and 5 cuttings of each transgenic genotype as the experimental units and three replications (blocks). Plants were harvested once in the year of transplanting. We took stand counts in the fall of 1996 and the spring of 1997 to determine survival. Plants were defoliated on June 28, July 28, and August 28, 1997, to determine dry-matter yields. The yields presented are the sums of the three harvests in 1997.

The 1997 field trial was also conducted at the Elora Research Station following protocols authorized by Plant Products Division, Agriculture and Agri-Food Canada (tests 97-UOG1-075-ALF02-177-ONO1-01, 97-UOG1-075-ALF03-236-ONO1, and 97-UOG1-075-ALF04-224-ONO1-01; approved May 12, 1997). Replicated plots were established on May 28, 1997, by transplanting rooted cuttings of each transgenic and control genotype in 1- × 2-m rectangular plots at 100 plants per plot. Each plot consisted of a population of independent transgenic plants for each construct. Plants were harvested twice in the year of transplanting on July 1 and September 2, 1997. On November 19, 1997, samples were dug from the field, rinsed, separated into taproot, crown, and leaves, and immediately submerged in liquid nitrogen. The samples remained in liquid nitrogen until they were ground in the laboratory and analyzed for SOD activity on native PAGE gels as described above. Whole samples were also dug from the field. Crowns and roots of these field-acclimated alfalfa plants were subjected to freezing temperatures by placing the plants in moist filter paper. The samples were frozen at −2°C overnight and then cooled at 2°C per hour to −6°C, −8°C, −10°C, −12°C, or −14°C. The frozen samples were thawed for 1 d at 2°C. The plants were separated into crowns, taproots, and crown buds, bisected, and assessed by viability staining with nitroblue tetrazolium as previously described (Tanino and McKersie, 1985). We repeated the test on December 2, 1998.

Statistical Analysis

We performed an analysis of variance using SAS for Windows, Proc GLM (version 6.11, SAS Institute, Cary, NC). Because of missing values in some experiments, we calculated least-square means. Significance was determined at the 5% level of probability.

RESULTS

Two clones of alfalfa were transformed with A. tumefaciens containing either the pMitSOD or pChlSOD binary vectors. Using PCR we screened the plants first for the presence of nos-nptII transgene-conferring kanamycin resistance. Approximately 90% of the regenerated plants scored positively in the screen. Only PCR-positive plants were transferred to the greenhouse for further study. Southern-blot analysis of eight transgenic plants confirmed that there were one or two full insertions of the tDNA in the chromosomes of each of the transgenic plants (data not shown).

We analyzed the transgenic plants from the clone N4 containing pMitSOD or pChlSOD for SOD expression using native PAGE (Fig. 1). The nontransgenic control N4 plant had three major SOD bands: a fast-moving chloroplastic form of Cu/Zn-SOD, a slower-moving cytosolic form of Cu/Zn-SOD, and a mitochondrial Mn-SOD. A small Fe-SOD peak was occasionally detected between the Mn-SOD and cytosolic Cu/Zn-SOD, but its activity was quite labile and not included in the calculations of total SOD activity. The transgenic plants had an additional Mn-SOD enzyme superimposed on the native Mn-SOD isozyme in the native gels. In the case of the pMitSOD, the two Mn-SOD forms were not resolved by PAGE, but in the case of the pChlSOD, two distinct Mn-SOD bands were apparent. The variance in the mobility of the mitochondrial and chloroplastic targeted forms of Mn-SOD possibly reflects differences in the cleavage site of the transit peptide or another posttranscriptional modification.

Figure 1.

Line scan of native PAGE gels for SOD activity in leaf extracts of transgenic alfalfa plants expressing pMitSOD or pChlSOD. Cu/Zn-SOD1 and Cu/Zn-SOD2 designate cytosolic and chloroplastic forms of SOD, respectively. Fe-SOD activity was observed as a small band only on some gels. The area of the Mn-SOD band was increased in both pMitSOD and pChlSOD compared with N4. A distinct double band was observed in Mn-SOD of pChlSOD but not pMitSOD.

We quantified the amount of each SOD isozyme in two ways. The area of each peak from the line scan was calculated and expressed relative to the total area of SOD activity in each lane to determine its proportion of total SOD activity. This method compensated for differences in the amount of the extract applied to the gel and for differences in the staining intensity among gels. However, the method assumed that Cu/Zn-SOD was not affected by expression of the Mn-SOD transgene. So, in some experiments we quantified the amount of each SOD isozyme by expressing its area relative to the area of an internal standard (bovine Cu/Zn-SOD) on the same gel to calculate specific activity (units per gram of protein). The amount of bovine Cu/Zn-SOD applied was linearly related to the area of the peak over the range used in these experiments (data not shown).

Based on native PAGE analysis, Mn-SOD activity increased by a variable amount among the independent transgenic plants containing either pMitSOD or pChlSOD (Table I). In approximately 25% of the transgenic plants from either transformation vector Mn-SOD activity lessoned or did not change. In the majority of the transgenic plants Mn-SOD activity increased less than 2-fold. In only a small proportion of the transgenic plants was Mn-SOD activity doubled relative to total SOD activity.

Table I.

Frequency of different Mn-SOD activity classes among independent transgenic alfalfa plants expressing pMitSOD or pChlSOD

Class	pMitSOD	pChlSOD
	no. of plants
≤Control	8  (23)	7  (29)
1%–50% Increase	14  (40)	9  (38)
51%–100% Increase	11  (31)	7  (29)
>100% Increase	2  (6)	1  (4)

Mn-SOD activity in leaf extracts from N4 was 19% of total SOD activity as determined by activity staining of native polyacrylamide gels. Values in parentheses represent the frequencies as percentages of the total number of transgenic plants.

We selected the highest expressing plants for more detailed analysis. In three pMitSOD and two pChlSOD plants, Mn-SOD activity increased in all shoot tissues sampled (Table II), but the increase was not uniform across tissues. The increase in Mn-SOD activity was least in apex, nodes, and stem and greatest in the petiole and blade leaf tissues. The response was similar in both vectors. The variation among tissues in the transgenic plants was possibly due to differences in the transcription of the cauliflower mosaic virus 35S promoter among the tissues, but posttranscriptional modifications of SOD activity were also likely to be important contributing factors.

Table II.

Mn-SOD activity as a percentage of total SOD activity in shoot tissues of transgenic alfalfa plants expressing pMitSOD or pChlSOD

Tissue	Control	pMitSOD	pChlSOD
	%
Apexa	19	32	39
Nodes	18	25	42
Leaf 1b	17	45	54
Leaf 2b	19	46	50
Leaf 3b	19	45	51
Leaf 4b	20	45	47
Petiole	17	52	47
Stem	18	40	37
lsd (0.05)c	NSd	2	4

Activities were measured on three high-expressing pMitSOD plants and two high-expressing pChlSOD plants.

^aApex includes the apical meristem and the young expanding leaf. 

^bLeaves 1 to 4 represent trifolidate leaf blades numbered sequentially from the first fully expanded leaf to the fourth. 

^clsd within a column at the 5% level of probability; n = 6 (control and pChlSOD) or 12 (pMitSOD). 

^dNS, Not significantly different according to analysis of variance. 

Bovine Cu/Zn-SOD was used as an internal standard, and the PAGE analysis showed that Mn-SOD specific activity also increased for both vectors (Table III). However, if the data were expressed as a percentage of total SOD, the increase was greater in pMitSOD than predicted. This discrepancy occurred because targeting the Mn-SOD to the chloroplasts reduced Cu/Zn SOD activity. The plants containing pMitSOD had the same activity of both forms of Cu/Zn-SOD as the N4 control; therefore, total SOD activity in leaf extracts increased. In contrast, targeting the Mn-SOD to the chloroplast caused an apparent feedback regulation, leading to lower chloroplastic and cytosolic Cu/Zn-SOD activity; therefore, total SOD activity actually lessened, even though Mn-SOD activity increased. Although the mechanism of this apparent regulation is unknown, there appeared to be similar regulation of the two Cu/Zn forms of the enzyme.

Table III.

Mn- and Cu/Zn-SOD activities in independent transgenic alfalfa plants expressing pMitSOD or pChlSOD

SOD Isozyme	Control	pMitSOD	pChlSOD
	units of SOD activity g−1 protein
Mn-SOD	426c  (18c)	1111a  (41b)	802b  (46a)
Cu/Zn-SOD1	883a  (38a)	759a  (24b)	371b  (19c)
Cu/Zn-SOD2	968a  (43a)	996a  (36b)	667b  (35b)
Total	2278b	2851a	1840c

Values are averaged across shoot tissues from the same experiment as described in Table II. Values in parentheses are percentages of activity of each individual form relative to the total SOD activity. Values within a row followed by the same letter are not significantly different according to Duncan's multiple range test at the 5% level of probability; n = 50 (control), 106 (pMitSOD), or 58 (pChlSOD).

We selected four highly expressing primary transgenic plants from each vector for growth analysis and stress-tolerance testing in the greenhouse. The plants were propagated with cuttings. We observed no variation in SOD activity among propagules of the same transgenic plant (data not shown). We tested cuttings in one experiment with PCR for nos-nptII and all remained positive (data not shown).

The transgenic plants and a nontransgenic control were grown to the early bud stage, defoliated, and allowed to grow new shoots as a measure of vigor. Three of the four pMitSOD plants tested had greater shoot dry-matter production than the control (Table IV). Two pChlSOD plants had less growth and two had more growth than the control.

Table IV.

Vigor of primary independent transgenic alfalfa plants expressing pMitSOD or pChlSOD

Control		pMitSOD		pChlSOD
Plant	Dry wt	Transformanta	Dry wt	Transformant	Dry wt
	mg	mg			mg
N4	635	1	639	1	695
		6	678	6	529
		9	720	7	555
		10	723	9	662

Shoot dry matter production in the greenhouse was determined at 28 d after defoliation. lsd at the 5% level of probability for dry weight determinations = 22, n = 28.

^aTransformant indicates an independent transformation event as confirmed by Southern-blot analysis. 

Freezing tolerances of leaf blades and fibrous roots were measured by electrolyte leakage. The plants were acclimated to three different stages and frozen to four different temperatures. Statistically significant main effects were observed for tissues, acclimation, and temperatures (Table V). Roots were consistently more freezing sensitive than leaves. Both leaves and roots had less injury with acclimation. Lower temperatures caused increasing amounts of injury. Small, statistically significant differences were observed among the transgenic plants and compared with the control plant. However, some transgenic plants had less injury than the nontransgenic N4 control, whereas others had more injury. Nonetheless, the most tolerant transgenic plant was only 1°C more freezing tolerant than the control, indicating that any improvement in freezing tolerance per se was minimal.

Table V.

Electrolyte leakage from leaves and roots after freezing of independent transgenic alfalfa plants expressing pMitSOD or pChlSOD

Main Treatment Effects	Electrolyte Leakage
	Leaf	Root
	% of total
Acclimation (weeks)
2	55	74
4	42	70
4 + 1a	NAb	56
lsd (0.05)	0.2 (n = 108)	0.2 (n = 108)
Temperature (°C)
−6	33	58
−7	49	67
−8	53	70
−9	58	73
lsd (0.05)	0.3 (n = 54)	0.3 (n = 81)
Genotype
Control N4  pMitSOD	51	68
N4 pMitSOD-1	53	67
N4 pMitSOD-6	52	66
N4 pMitSOD-9	48	66
N4 pMitSOD-10	47	66
pChlSOD
N4 pChlSOD-1	47	64
N4 pChlSOD-6	44	66
N4 pChlSOD-7	46	70
N4 pChlSOD-9	45	68
lsd (0.05)	0.6 (n = 24)	0.5 (n = 36)

The experiment was factorial with a split plot design of three acclimation stages (main plot-different weeks) × 7 plants × 4 temperatures × 3 replications (different days). Leaf and root samples were analyzed separately. Values for the main treatment effects are averaged across other treatments in the factorial experiment.

^aPlants were acclimated at 2°C for 4 weeks, frozen at −2°C, and then acclimated at 2°C for 1 more week. Leaves were injured by the −2°C acclimation treatment. 

^bNA, Data not available. 

Independent transgenic plants were selected from the N4 group for the field trial if they had increased SOD activity; N4 plants with lower SOD activity were not included. The other S4 group of plants was not tested in the greenhouse and all available independent transgenic plants were placed in the field. All plants had good survival in the year of transplanting and entered the winter of 1996/1997 with close to 100% survival. The location of the field trial experienced a relatively harsh winter in 1996/1997 because of excessively wet soil conditions. This site was chosen because the permit for the transgenic field trial dictated that the experimental alfalfa plants had to be isolated by 20 m from any other alfalfa and because less than 100% survival of the nontransgenic clones was desired. Survival in spring 1997 was less than commonly observed with these parental clones in other tests. The average survival of the transgenic plants was higher than that of the nontransgenic controls (Table VI). The N4 group of transgenic plants had higher survival and vigor in the field than the S4 group, but the two vectors pMitSOD and pChlSOD were similar in both.

Table VI.

Average survival and yield of independent transgenic alfalfa plants expressing pMitSOD or pChlSOD in spaced plantings

Alfalfa Clone	Vector	No. of Independent Transgenics	Survival	Yield
			%	g/plant
N4	Control	1	53	34
	pMitSOD	10	69	50
	pChlSOD	9	69	36
S4	Control	1	29	15
	pMitSOD	13	50	19
	pChlSOD	1	67	52

Two parent clones designated as N4 and S4 were transformed with the pMitSOD or pChlSOD vectors. Values are the averages across all independent transgenic plants within a vector with three replications of 5 plants per independent transgenic plant and 15 plants per control. Survival and yield of the transgenic plants were significantly greater than the nontransgenic control plants at P = 0.0243 and P = 0.0159, respectively, according to an orthogonal contrast.

The genetic potential of the plant and the quantity of stored carbohydrate reserves determine the first-cut herbage yield, but winter injury to the root, crown, or crown buds may reduce the yield. The average yield of the transgenic plants expressing pMitSOD was higher than that of the N4 and S4 nontransgenic controls (Table VI). The average yield of those expressing pChlSOD was the same as N4 and significantly greater than S4, although in the latter case there was only one transgenic plant tested.

We observed more variation of both survival and yield within each group of transgenic plants than between parent clones or vectors. The N4 nontransgenic control had 53% survival, and S4 had 29%. Most transgenic plants had improved survival. Some individual transgenic plants in both pMitSOD and pChlSOD had 100% survival, but there were other individual plants with less survival than the nontransgenic controls in both parental groups and with both vectors (Table VII). Similarly, although both pMitSOD and pChlSOD had individuals with dramatic improvements in yield, approaching twice the yield of the N4 nontransgenic control, other individuals had lower yields than the nontransgenic controls (Table VIII).

Table VII.

Frequency of plants in survival classes of independent transgenic alfalfa plants expressing pMitSOD or pChlSOD in spaced plantings

Survival Class	N4			S4
	Control	pMitSOD	pChlSOD	Control	pMitSOD	pChlSOD
%
0–20	–a	–	–	–	23	–
21–40	–	10	–	100	8	–
41–60	100	40	44	–	38	–
61–80	–	20	33	–	23	100
81–100	–	30	22	–	8	–

Values are the percentages of the independent transgenic plants in each survival class. No. of plants and average values are shown in Table VI.

^a–, No plants were observed in this class.  

Table VIII.

Frequency of yield classes of independent transgenic alfalfa plants expressing pMitSOD or pChlSOD in spaced plantings

Yield Class	N4			S4
	Control	pMitSOD	pChlSOD	Control	pMitSOD	pChlSOD
g	%
0–10	–a	–	–	–	23	–
11–20	–	10	22	100	23	–
21–30	–	10	44	–	38	–
31–40	100	30	–	–	15	–
41–50	–	–	11	–	–	–
51–60		10	11	–	–	100
>60	–	40	11	–	–	–

Values are the percentages of the independent transgenic plants in each yield class. No. of plants and average values are shown in Table VI.

^a–, No plants were observed in this class. 

This experiment tested only primary transgenic plants because it was necessary to identify the elite transformation event for future genetic and physiological evaluation. Therefore, it is possible that the differences in winter survival among these transgenic plants were related to the tissue culture methods used to create them. However, an adjacent field experiment compared the effects of the three vectors containing the yeast suc2 gene and did not identify any with greater survival or yield than the control (D. Vadnais, unpublished data). Therefore, improved winter survival was not a common response observed with this transformation system. A comparison of the winter survival of progeny of the selected primary transgenic plants segregating for the SOD transgene is in progress.

To determine whether these field results were due to differences in SOD expression patterns between the greenhouse and the field, a random sample of N4 transformed and control plants was dug from the field and analyzed using native PAGE (Table IX). Results were similar to previous results from the greenhouse-grown plants. Mn-SOD (expressed as a percentage of total SOD) increased significantly in both pMit-SOD and pChl-SOD vectors (P = 0.03). On average, 53% of total SOD activity was due to cytosolic Cu/Zn-SOD, and 23% was due to chloroplastic Cu/Zn SOD. Chloroplastic Cu/Zn-SOD had lower activity in both pMit-SOD and pChl-SOD plants (data not shown).

Table IX.

Mn-SOD activity as a percentage of total SOD activity in tissues of transgenic alfalfa plants expressing pMitSOD or pChlSOD

Tissue	Control	pMitSOD	pChlSOD
	%
Leaf	22	17	29
Crown bud	16	21	28
Crown	21	33	31
Taproot	14	27	25
Mean	18	26	28

Activities were measured on plants taken from the field on November 19, 1997. lsd at the 5% level of probability = 14; n = 3. Average Mn-SOD activity = 1398 units g⁻¹ protein.

The crowns and roots of field-acclimated plants were sampled from the 1997 field trial in November and again in December, and then they were subjected to freezing temperatures. Viability was determined by vital staining with tetrazolium. In the samples removed from the field on December 2, 1998, the cortex of the taproots lost viability below −8°C (no red staining as shown in Fig. 2). At −10°C the pith of the crowns was affected and the first signs of injury appeared in the axis and innermost scales of the crown buds. At −14°C the bud axis and inner scales were dead. The vascular cylinder of the taproots was almost white, although the endodermis still stained red. The vascular system and cortex of the crown also remained viable. When the same test was performed 2 weeks before, the extent of injury in all tissues was much greater. In the crowns the vascular system was the least vulnerable, and in the crown buds the very tip of the primordia survived −14°C but the vascular tissues were no longer viable. We found no differences in the sites or pattern of freezing injury between the nontransgenic and any of the transgenic plants (data not shown). At −10°C and −14°C we observed the same extent of freezing injury in all of the transgenic plants and in the control (nontransgenic) plants.

Figure 2.

Viability staining of the roots and crown of alfalfa. Plants were dug from field plots on December 2, 1997, frozen to −6°C, −10°C, and −14°C, and stained with nitroblue tetrazolium to detect viability. Cells stained red have active respiration and are considered alive. Those stained white have been injured. There is essentially no difference in the injury seen between control and SOD-transgenic plants at any freezing temperature examined.

DISCUSSION

Transgenic alfalfa plants constitutively expressing a Mn-SOD cDNA had greater Mn-SOD activity, significantly greater survival in the field after one winter, and greater total shoot dry-matter production (yield) in the field than the nontransgenic control plants. This effect has now been observed in three alfalfa clones (N4 and S4 in this study and RA3 in a previous study [McKersie et al., 1993, 1996]) with two SOD vectors having different subcellular targeting sequences. However, this study shows that there was a large amount of variation among transgenic plants regenerated from a single vector in a single clone. This variation occurred in all of the parameters measured, including Mn-SOD activity, yield in the greenhouse and in the field, and winter survival. The variation among plants that had been transformed with the same vector was greater than that due to subcellular targeting of SOD or due to the alfalfa clone into which the tDNA was inserted. We now see that our earlier experiments with RA3 were fortuitous in showing a positive effect because it is clear from these data that there was a possibility of selecting individual transgenic plants with less hardiness, the same hardiness, or improved hardiness from the same set of transgenic plants.

These data indicate that we must be cautious in interpreting experiments that measure stress-tolerance effects in transgenic plants using only a few independent transgenic plants from each vector, such as in our previous study (McKersie et al., 1996), because conflicting results are likely. For example, in the case of SOD overexpression, Tepperman and Dunsmuir (1990), Pitcher et al. (1991), and Payton et al. (1997) found no improvement in tolerance to oxidative and related stresses, whereas Bowler et al. (1991), Gupta et al. (1993a, 1993b), Perl et al. (1993), McKersie et al. (1993, 1996), and Van Camp et al. (1994, 1996) found significant improvements. Several studies (Pitcher et al., 1991; Gupta et al., 1993a, 1993b) evaluated only two transgenic plants. The present data suggest that a reliable prediction of a general trend should require more than 20 independent transgenic alfalfa plants.

The alfalfa plants that we produced varied considerably in SOD activity, so we attempted to select plants in the greenhouse to reduce the number of primary transgenic plants evaluated in the field and to speed the selection of plants for cross-pollination and seed production. Prescreening the transgenic plants for SOD activity, vigor, or freezing tolerance in the greenhouse was not effective at identifying individual transgenic plants with improved field performance. One may have anticipated this because survival of the crowns and roots determine winter survival of alfalfa, and we measured SOD activity in shoots in the greenhouse. Although the greenhouse is a very different environment than the field, we assumed that the cauliflower mosaic virus 35S promoter would be constitutive and that therefore individuals with high expression in one environment and one tissue would have correspondingly high expression in another. Either this assumption was not valid or other factors interacted with SOD activity to modify performance.

Shoot growth in the greenhouse was also not a reliable prediction of growth in the field. Comparing the yields of transgenic plants common to Tables IV and VIII by regression analysis gave r² = 0.32. A greenhouse screen for vigor may have eliminated a few poorly performing plants, but it could not identify the highly performing plants.

Freezing tolerance increased by only 1°C in the best transgenic plant when the plants acclimated to controlled environments and when we measured viability by electrolyte leakage. When the crowns and taproots of field-acclimated plants were frozen, there was no apparent difference in freezing tolerance between transgenic and control plants, if we measured viability by tetrazolium staining. Therefore, these studies do not indicate that overexpression of SOD changed the primary site of freezing injury in alfalfa. Yet, we did observe differences in winter survival among the plants. Previously, we had observed differences when viability was measured as the amount of shoot regrowth from crowns after a freezing stress (McKersie et al., 1993). Additional SOD activity may have improved the cellular repair mechanisms, thereby allowing the transgenic plant to better recover from freezing injury. Perhaps cellular viability was not a good indicator of whole-plant viability in this instance. In addition, freezing may not have been the most appropriate laboratory stress test, and better outcomes may have resulted from ice-encasement, flooding, or long-duration freezing tests. Unfortunately, despite the current regulatory requirements for permits to test transgenic plants in the field, field testing remains the most suitable measure for winter survival and yield potential.

Abbreviation:

SOD

superoxide dismutase