This article was downloaded by: [Dr Manjit S. Kang]On: 09 August 2015, At: 03:42Publisher: Taylor & FrancisInforma Ltd Registered in England and Wales Registered Number: 1072954 Registeredoffice: 5 Howick Place, London, SW1P 1WG

Journal of Crop ImprovementPublication details, including instructions for authors andsubscription information:http://www.tandfonline.com/loi/wcim20

Transgene Transfer to United StatesCommercial Rice Cultivars viaConventional Breeding TechniquesQ. M. Shao a , M. C. Rush a , M. S. Wu c , D. E. Groth d , M. S. Kang b

& S. D. Linscombe da Department of Plant Pathology & Crop Physiology , Louisiana StateUniversity Agricultural Center , Baton Rouge, LA, 70803, USAb Department of Agronomy , Louisiana State University AgriculturalCenter , Baton Rouge, LA, 70803, USAc The Institute of Plant Protection, Chinese Academy of AgriculturalScience , 2 West Yuanmingyuan Road, Beijing, 100094, Chinad Rice Research Station , Louisiana State University AgriculturalCenter , Crowley, LA, 70527, USAPublished online: 21 Sep 2008.

To cite this article: Q. M. Shao , M. C. Rush , M. S. Wu , D. E. Groth , M. S. Kang & S. D. Linscombe(2006) Transgene Transfer to United States Commercial Rice Cultivars via Conventional BreedingTechniques, Journal of Crop Improvement, 16:1-2, 49-65, DOI: 10.1300/J411v16n01_04

To link to this article: http://dx.doi.org/10.1300/J411v16n01_04

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the“Content”) contained in the publications on our platform. However, Taylor & Francis,our agents, and our licensors make no representations or warranties whatsoever as tothe accuracy, completeness, or suitability for any purpose of the Content. Any opinionsand views expressed in this publication are the opinions and views of the authors,and are not the views of or endorsed by Taylor & Francis. The accuracy of the Contentshould not be relied upon and should be independently verified with primary sourcesof information. Taylor and Francis shall not be liable for any losses, actions, claims,proceedings, demands, costs, expenses, damages, and other liabilities whatsoever orhowsoever caused arising directly or indirectly in connection with, in relation to or arisingout of the use of the Content.

This article may be used for research, teaching, and private study purposes. Anysubstantial or systematic reproduction, redistribution, reselling, loan, sub-licensing,systematic supply, or distribution in any form to anyone is expressly forbidden. Terms &

Conditions of access and use can be found at http://www.tandfonline.com/page/terms-and-conditions

Dow

nloa

ded

by [

Dr

Man

jit S

. Kan

g] a

t 03:

42 0

9 A

ugus

t 201

5

Transgene Transfer to United StatesCommercial Rice Cultivars

via Conventional Breeding Techniques

Q. M. ShaoM. C. RushM. S. Wu

D. E. GrothM. S. Kang

S. D. Linscombe

ABSTRACT. Breeding programs can benefit from transfer of a foreigngene from one transgenic plant to commercial cultivars through continu-ous backcrossing, especially to cultivars in which it is difficult to trans-fer a foreign gene directly through biotechnology. In this study, twohomogeneous transgenic plants, T-28 and T-64 (from Taipei 309), andone homogeneous transgenic plant, N-84 (from Nipponbare), were usedas donors of Bar gene. Commercial cultivars, Cypress and Laffite, wereused as the recurrent parents. Three to five backcrosses were made usingthe transgenic plants as donor parents and the commercial cultivars asrecurrent parent. The results from selected progeny rows, and two-yearsof yield tests with selected lines, indicated that the target Bar gene couldbe transferred to lines from homozygous transformants in 2-3 years of

Q. M. Shao and M. C. Rush are affiliated with the Department of Plant Pathology &Crop Physiology, and M. S. Kang is affiliated with the Department of Agronomy; allare at Louisiana State University Agricultural Center, Baton Rouge, LA 70803.

Present address for Q. M. Shao: Bayer Crop Science, 269 Quail Dr, Woodland, CA95695.

M. S. Wu is affiliated with The Institute of Plant Protection, Chinese Academy ofAgricultural Science, 2 West Yuanmingyuan Road, Beijing 100094, China.

D. E. Groth and S. D. Linscombe are affiliated with Rice Research Station, Louisi-ana State University Agricultural Center, Crowley, LA 70527.

Journal of Crop Improvement, Vol. 16(1/2) (#31/32) 2006Available online at http://www.haworthpress.com/web/JCRIP

© 2006 by The Haworth Press, Inc. All rights reserved.doi:10.1300/J411v16n01_04 49

Dow

nloa

ded

by [

Dr

Man

jit S

. Kan

g] a

t 03:

42 0

9 A

ugus

t 201

5

50 JOURNAL OF CROP IMPROVEMENT

backcrossing, giving lines similar to the recurrent parents in phenotypeand yield potential. [Article copies available for a fee from The Haworth Doc-ument Delivery Service: 1-800-HAWORTH. E-mail address: <docdelivery@haworthpress.com> Website: <http://www.HaworthPress.com> © 2006 by TheHaworth Press, Inc. All rights reserved.]

KEYWORDS. Oryza sativa, transformation, Bar gene, Liberty herbi-cide, backcrossing

INTRODUCTION

The purpose of backcross breeding is to improve an otherwise desir-able cultivar (recurrent parent) by introducing into it a gene or a set oflinked genes from a donor parent (Fehr et al., 1987). Several genera-tions of backcrossing is needed to transfer a major gene from one plantto another plant (Allard, 1960; Nelson, 1973). Genes for sterility andfertility restoration have been successfully transferred from wild rice, orother sources, into hybrid rice cultivars or elite breeding lines in China(Yan et al., 1991). Dominant genes for resistance to bacterial leaf blight,caused by Xanthomonas oryzae pv. oryzae, also were transferred fromTetep, DV-78, and other sources into commercial cultivars (Shao et al.,1991; Wan et al., 1991). The effects of major resistance genes on ricebacterial leaf blight were modified by other genes through continuousbackcrossing (Shao et al., 1991, 1992). The results of a backcross recip-rocal monosomic analysis of Fusarium head blight resistance, using thehighly resistant Hungarian winter wheat (Triticum aestivum L.) line‘U-136.1’ and the highly susceptible cultivar ‘Hobbit-sib’ indicated thatthe families containing U-136.1 chromosomes 6B, 5A, 6D, 1B, and 4Bhad a visually reduced infection spread compared with lines that had the‘Hobbit-sib’ chromosome (Buerstmayr et al., 1999).

Transgenic Taipei 309 and Nipponbare plants, which were tested inour research, are japonica-type rice, possessing pubescence, mediumgrain, and other undesirable agronomic characteristics. These transgenicplants containing the Liberty herbicide resistance (Bar) gene cannot beused directly as cultivars by growers in the rice-producing areas of theUnited States. There were problems in directly transferring the foreigngene into current rice cultivars, for example, the cultivars were hard toregenerate in tissue culture; many plants obtained through tissue culturewere sterile, and somaclonal variation often occurred in regeneratedplants originating from transformation (Xie et al., 1995; Christou et al.,

Dow

nloa

ded

by [

Dr

Man

jit S

. Kan

g] a

t 03:

42 0

9 A

ugus

t 201

5

1997). Therefore, it is desirable to transfer the Bar gene from trans-formed plants into commercial cultivars through repeated backcrosses.Reports exist on improvement, via backcrossing, of characteristics ofrice cultivars in which they were deficient (Shao et al., 1991; Wan et al.,1991; Yan et al., 1991; Buerstmayr et al., 1999; Zemetra et al., 1999).There are, however, no reports related to the transfer of foreign genesfrom transformed plants to cultivars through repeated backcrosses. Thepurpose of this research was to examine the utility and efficacy of usingbackcross methodology to transfer alien genes from transformed plantsinto commercial lines or cultivars.

MATERIALS AND METHODS

Crossing and Line Selection from Progeny Rows

The transgenic lines T-28 and T-64 from Taipei 309, and N-84 fromNipponbare were used as Liberty resistance sources and donor parents.Commercial rice cultivars commonly grown in the southern UnitedStates, including Cypress and Lafitte, were used as the recurrent par-ents. These cultivars are naturally susceptible to Liberty herbicide. Seedof all the parents were germinated in the greenhouse in the spring of 1996and were manually transplanted 30 days after seeding into the field atthe Rice Research Station (RRS) of the Louisiana State University Ag-ricultural Center in Crowley, LA. The field received a preplant applica-tion of 14-42-42 (N-P2O5-K2O) fertilizer and topdressing of 494.2 kgha–1 urea. The experimental area was treated with herbicide (propanil andLondax) and Icon insecticide. Standard agronomic practices were usedthroughout the rice-growing period.

Three to five plants per parent were removed from the field and trans-planted into the greenhouse at the boot or early heading stage. In thesummer of 1996, panicles on the plants that were heading but had notflowered were selected to make crosses using hot water treatment. Seedswere harvested 3 weeks after pollination. The harvested seeds were driedin an oven at 41�C for 3 days.

Seeds from the crosses were germinated in the laboratory, and theseedlings were transplanted into the greenhouse 10 days after germina-tion during the winter of 1996-1997. Thirty days after transplanting, allprogeny from each cross were treated with 0.228 kg a.i. ha�1 hygromycinB solution by cutting leaf tips submerged into the solution. Lesions onthe treated leaf tips were measured 7 days later. Liberty herbicide was

Shao et al. 51

Dow

nloa

ded

by [

Dr

Man

jit S

. Kan

g] a

t 03:

42 0

9 A

ugus

t 201

5

sprayed at a rate of 0.414 kg a.i. ha–1 onto the plants in the greenhouse toevaluate their resistance. The surviving plants (double-resistant) wereselected for use as parents to make backcrosses with the recurrent par-ents. BC1 seeds were germinated in the laboratory and seedlings weretransplanted into the greenhouse 10 days after germination. Seedlingswere transplanted into the field by hand at the RRS in Crowley, LA aftergrowing in the greenhouse for 25 days during the 1998 growing season.Standard agronomic practices were used in the tests. Hygromycin sensi-tivity was tested with 0.228 kg a.i. ha–1 hygromycin B solution per leaftip cutting 30 days after the seedlings were transplanted into the field.Lesion length data were collected 7 days after treatment. Liberty herbi-cide at a rate of 2.10 kg a.i. ha–1 was sprayed onto the plants 7 days afterthe hygromycin test. The number of plants killed or surviving and sensi-tivity of the surviving plants to hygromycin were determined. Survivingdouble-resistant plants, which were most similar to the recurrent par-ents, were selected and backcrossed again to the recurrent parents. Thesame method was used in the greenhouse and field tests through the BC5generation from 1996 to 2000.

A total of 50 lines was selected after the repeated backcrossing in2000. All progeny rows were planted at the RRS in tiers of 91 cm rowswith 25 cm spacing between rows using a Hege 90 Series Drill Planterwith 12-cell magazines. Each line was planted in five single rows withtwo replications. The respective recurrent and donor parents were used ascontrols. All experimental areas were treated with herbicide (propaniland Londax) and Icon insecticide. The fertilizer application and agro-nomic practices were the same as mentioned previously. One replicationwas treated with Liberty at a rate of 2.10 kg a.i. ha–1 50 days after seeding.The other replication was used as the untreated control. Major agronomiccharacteristics, including plant height, days from planting to heading, andresistance or susceptibility to Liberty and hygromycin B for each line,were measured. Twelve lines were harvested from the test and used as en-tries in yield tests during the 2001 and 2002 growing seasons.

Yield Tests of Selected Lines from Liberty-ResistantProgeny Rows

Twelve homozygous lines selected from different backcross genera-tions were used as entries and yield tests were conducted in the field atthe RRS in 2001 and 2002. The corresponding recurrent parents of theselines, Cypress and Lafitte, were planted as controls. The lines and con-trols were planted in 122 × 488 cm2 subplots with 36 cm alleys between

52 JOURNAL OF CROP IMPROVEMENT

Dow

nloa

ded

by [

Dr

Man

jit S

. Kan

g] a

t 03:

42 0

9 A

ugus

t 201

5

plots and 122 cm alleys between tiers of plots. The plots were seededon April 30, 2001 and on April 23, 2002. A split plot on randomizedcomplete block (RCB) with three replications was used. The main plotswere lines or cultivars and the sub-plots were Liberty treatments (withor without Liberty). The field received a preplant application of 24-72-72(N-P2O5-K2O) fertilizer. All plots received a pre-flood application of62.2 kg ha–1 of ammonium sulfate fertilizer and an additional 113.6 kgha–1 top-dressing with urea. Standard agronomic practices were fol-lowed. The soil was silt loam (pH 6.0, clay 12%, silt 71%, sand 17%,CEC 9.4 mole kg–1). Each plot was 122 × 488 cm2. Liberty at the rate of2.10 kg a.i. ha–1 was applied at the four-leaf stage of growth. Data werecollected 3 days after Liberty application. The number of days fromplanting to heading was recorded on the basis of 50% panicle emer-gence in each plot. Plant height of mature plants was measured from thecrown (soil line) to the tip of the highest panicle. The plots were har-vested on August 20, 2001 and on August 26, 2002.

Grain for testing milling quality was harvested by hand from each plot.After the grain was dried and cleaned, 125 g of grain was taken from eachplot and used as milling samples. Milling quality was tested at the GrainQuality Laboratory at the RRS and the standard procedure was fol-lowed.

Seeds from each entry were harvested by hand, mixed well, and driedfor 3 days in an oven at 41�C. After cleaning the dried seeds, 10 grainswere randomly selected, and grain length and width measurements wererecorded in millimeters. Seed weight of 500 randomly selected grainswas recorded in grams. Three replications per plot were used.

Gene Detection from the Yield Test via PolymeraseChain Reaction (PCR)

A total of 8-10 leaves from the yield-test entries were collected 10 daysafter Liberty application. The total genomic DNA from all of the entrieswas isolated and purified. PCR was used to amplify the target Bar gene.Rice genome DNA was prepared based on the protocol for molecularcloning (Sambrook et al., 1989). One pair of primers was used for Bargene amplification (polymerase chain reaction, PCR). Another primerpair was used for hpt gene amplification (forward: 5-AGTTCGACAGCGTCTCCGA-3; reverse: 5-ATTCCTTTGCCCTCGGACGA-3).Genomic DNA isolated from the tested material was used as templates.The components of PCR reaction included 5 µl of 10 × PCR buffer, 4 µl

Shao et al. 53

Dow

nloa

ded

by [

Dr

Man

jit S

. Kan

g] a

t 03:

42 0

9 A

ugus

t 201

5

of 25 mM MgCl2, 1 µl of 10 mM dNTPs, 1 µl of 10 µM forward primer,1 µl of 10 µM reverse primer, 2 units of Taq DNA polymerase, and fi-nally H2O up to 50 µl was added. These chemicals were gently mixed. Aquantity of 40 µl of light mineral oil was placed on the top of the mixturein a test-tube cap. The tubes were placed in adjacent wells of the thermalcycler (TECHNE, UK). Each cycle was composed of 94�C 50 secondsfor template denaturation, 54�C 50 seconds for primer annealing to tem-plate, and 72�C 2 minutes 30 seconds for DNA elongation. Thirty-fivecycles were run each time. The samples were run on 1% agarose gel todetect the results after gene amplification. SAS package 6.12 for Win-dows® was used to analyze all of the data (SAS Institute, 2000).

RESULTS

Changes in Major Agronomic Characteristicsof Progeny Rows Selected from Backcrosses

Transgenic plants used as donors for Liberty resistance in the currentstudies had characteristics that were controlled by a single or a fewdominant transgenes (Rush, 1999). The Liberty resistance in transgenicplants, T-28, T-64, and N-84, was controlled by the Bar gene. When Cy-press and Lafitte were used as the recurrent parents, data were collectedafter Liberty applications at the four-leaf seedling stage in the green-house or field tests from 1997 to 2000. The segregation ratio, which wasbased on the number of plants resistant to Liberty to the number ofplants susceptible to Liberty from the BC1 to BC5 generation, was 1:1with P-values ranging from 0.10 to 0.995 (Table 1). This corroboratedthe results on inheritance of the Bar gene in previous studies. The Bargene was inherited as a single dominant gene. When Lafitte was used asthe recurrent parent to make crosses with N-84, the agronomic charac-teristics of progeny changed quickly. After the second backcross, theplants selected for the repeated backcrosses were already similar to therecurrent parent, Lafitte, for several agronomic characters, such as plantheight (97.8 cm), days from planting to heading (78 days), mean paniclelength (21.3 cm), number of spikelets per panicle (145.8), and seed-setrates (89.6%) (Table 2). Plants were significantly taller than the donorparent (81.5 cm) and closer to the recurrent parent (98.5 cm). The days-to-heading, panicle length, and number of spikelets per panicle of theplants selected were similar to Lafitte’s corresponding characteristics

54 JOURNAL OF CROP IMPROVEMENT

Dow

nloa

ded

by [

Dr

Man

jit S

. Kan

g] a

t 03:

42 0

9 A

ugus

t 201

5

(Table 2). The seed set did not change. This was because of the similarityof the seed-set rates for both the donor and recurrent parent (Table 2).

When Cypress was used as the recurrent parent in a cross with N-84,some characteristics, such as plant height and days-to-heading, changedgradually, and were intermediate between the donor and recurrentparents in BC2 (Table 2). Plant height after five repeated backcrosseswas 97.2 cm, which was about the same as that of the recurrent parent(97.3 cm). The days-to-heading of the plants selected at the secondgeneration was 75 days, which was in between the donor and recurrentparents. The days-to-heading of the plants selected in the fifth generationwas 83 days, which was similar to the days-to-heading of the recurrentparent Cypress (81 days). Some of the agronomic characters of theprogeny, such as panicle length, number of spikelets per panicle, andseed-set rates, changed rapidly towards the characteristics of therecurrent parent. These characteristics were similar to Cypress, evenin the second backcross generation progeny (Table 2).

Shao et al. 55

TABLE 1. Segregation of the Bar gene in greenhouse or field tests in BC1through BC5 generation progeny following repeated backcrossing during1997-2000.

R† S‡ ExpectedRatio (R/S)

P-Value R S ExpectedRatio (R/S)

P-Value*

BC1 BC2T-28/Cypress 8 9 1:1 0.75-0.90 21 16 1:1 0.25-0.50T-64/Cypress 1 0 1:1 0.75-0.90 10 10 1:1 0.99-1.00N-84/Cypress 2 3 1:1 0.50-0.75 6 5 1:1 0.75-0.90N-84/Lafitte 3 2 1:1 0.50-0.75 14 8 1:1 0.10-0.25

BC3 BC4T-28/Cypress 9 6 1:1 0.25-0.50 2 0 1:1 0.10-0.25T-64/Cypress 5 8 1:1 0.25-0.50 4 1 1:1 0.10-0.25N-84/Cypress 3 6 1:1 0.25-0.50 3 6 1:1 0.25-0.50N-84/Lafitte 3 8 1:1 0.10-0.25 5 8 1:1 0.25-0.50

BC5T-28/Cypress 43 39 1:1 0.90-0.95T-64/Cypress ND NDw 1:1 – BlankN-84/Cypress 3 4 1:1 0.50-0.75N-84/Lafitte 12 15 1:1 0.50-0.75†R refers to the number of plants that were resistant to Liberty herbicide.‡S refers to the number of plants that were susceptible to Liberty herbicide.P-value is the smallest level of significance for which we would reject the null hypothesis.wND refers to no data available.

Dow

nloa

ded

by [

Dr

Man

jit S

. Kan

g] a

t 03:

42 0

9 A

ugus

t 201

5

Changes in the Agronomic Characteristics of Lines Selectedfrom Backcrosses in Yield Tests

Comparison with Backcross Lines Developed Using Cypressas the Recurrent Parent

Seven backcross lines, with Cypress as the recurrent parent, wereused in the 2001 and 2002 yield tests. Their major agronomic character-istics, including plant height, days-to-heading, yield per hectare, andgrain shape varied. Lines 230251, 230221, and 230413 had the sameyield potential as Cypress. Experimental line 230251 had the highestmean yield across the two years of testing (8271.8 kg ha–1). This washigher than mean yield of Cypress (7995.0 kg ha–1), although the differ-ence was not significant. Three of seven of the tested entries yielded sig-nificantly lower than Cypress. Experimental line 230420 had the lowestmean yield in the tests (6454.9 kg ha–1) (Table 3).

There were significant differences in plant height among the sevenentries and Cypress (Table 3). Mean height ranged from 81.3 cm to

56 JOURNAL OF CROP IMPROVEMENT

TABLE 2. Comparison of major agronomic characteristics between lines,which were generated from repeated backcrosses when Lafitte or cypress wasused as the recurrent parent and N-84 was used as the donor parent, and do-nor and recurrent parents in the field test of 2000.

Generation Mean PlantHeight (cm)

Mean Numberof Days-to-

Heading (days)

Mean PanicleLength (cm)

Mean Number ofSpikelets per

Panicle

Seed-SetRate (%)†

LafitteN-84 81.5 1.8‡ 71 16.5 1.5‡ 85.9 10.2‡ 88.5 2.4‡

Lafitte 98.5 2.9 81 21.4 1.7 169.8 10.1 85.9 2.8BC2 97.8 1.9 78 21.3 1.5 145.8 15.7 89.6 2.4BC3 95.8 1.9 79 21.6 1.3 181.2 15.1 91.6 2.2BC4 101.2 2.8 82 20.5 1.2 183.4 8.6 82.1 3.3BC5 100.1 1.6 80 20.9 1.2 169.2 12.1 83.4 2.1

CypressN-84 81.5 1.8‡ 71 16.5 1.5‡ 85.9 10.2‡ 88.5 2.4‡

Cypress 97.3 1.7 81 22.8 0.8 150.4 12.1 81.3 6.2BC2 86.8 1.9 75 22.1 1.3 160.8 6.1 80.4 3.1BC3 99.4 1.5 84 22.8 0.6 166.6 7.9 78.4 3.2BC4 98.7 2.0 83 22.5 0.9 163.4 6.3 80.2 2.2BC5 97.2 2.1 83 22.6 0.6 152.3 8.3 80.5 4.7†Seed setting rate refers to the number of filled seeds over total spikelets.‡Data include the average value for the characteristic plus standard deviation.

Dow

nloa

ded

by [

Dr

Man

jit S

. Kan

g] a

t 03:

42 0

9 A

ugus

t 201

5

100.2 cm. The Cypress control had a mean height of 92.8 cm, whichwas significantly shorter than line 230251 (100.2 cm). However, Cy-press was significantly taller than lines 230408 (88.3 cm) and 230413(87.0 cm). There were no significant differences in plant height betweenCypress and lines 230221 (91.8 cm) and 230281 (94.8 cm) (Table 3).All the line heights would be acceptable for commercial production.

The growth period from planting to heading is another importantcharacteristic in rice production. It varied from 85.4 days to 89.5 daysacross the Cypress-derived lines in the two years of field-testing (Table 3).Experimental line 230413 had a mean of 85.4 days from planting toheading, which was significantly smaller than the 87.5 days needed byCypress, the recurrent parent. Other lines, such as 230081 (87.7 days),230251 (88.2 days), and 230281 (88.2), were similar to Cypress. Exper-imental line 230420 needed about 89 days from planting to heading,which was longer than that of Cypress, but the difference was not signif-icant (Table 3).

The major agronomic characteristics varied significantly in differentyears. In 2001, the mean yield of the entries was 6158.0 kg ha–1, whichwas significantly lower than that in 2002 (7043.0 kg ha–1). The meanplant height was 91.5 cm in 2001, which was significantly greater thanthat in 2002 (87.7 cm). The mean days-to-heading in 2001 was 83.9days, which was significantly shorter than that in 2002 (91.7 days) (datanot shown). This was probably related to the weather during the rice-growing seasons. Year 2002 summer was cooler, with more rain, than2001 summer.

Shao et al. 57

TABLE 3. Comparison of yield, plant height, and days-to-heading betweenbackcross lines and their recurrent parent Cypress when no Liberty was ap-plied in the field tests in 2001 and 2002 combined.†

Entries Mean Yield(kg/ha)

Mean Plant Height(cm)

Mean Days-to-Heading(days)

Cypress 7995.0ab 92.8b 87.5ab230081 5626.0d 81.8d 87.7ab230221 7297.2abc 91.8b 89.0a230251 8271.8a 100.2a 88.2ab230281 7163.9abc 94.8b 88.2ab230408 6463.7cd 88.3c 86.2bc230413 6624.4bcd 87.0c 85.4c230420 6454.9cd 82.8d 89.5a†Means followed by different letters were significantly different at p = 0.05.

Dow

nloa

ded

by [

Dr

Man

jit S

. Kan

g] a

t 03:

42 0

9 A

ugus

t 201

5

Grain shape is a very important characteristic to growers and to the riceindustry. The mean grain length for backcross lines 230221, 230251, and230420 ranged from 9.3 to 9.6 mm (Table 4). This was about the samelength as that of the recurrent parent Cypress (9.4 mm grain–1). The meangrain length for lines 230281 and 230403 was about 9.00 mm grain–1,which was significantly shorter than that of Cypress. The mean grainlengths for lines 230081 and 230413 were the shortest in this group (Ta-ble 4). The mean grain width of lines 230251 and 230413 was 2.7 mm and3.0 mm, respectively, which was significantly greater than that of Cypress(2.5 mm). There were no significant differences in grain width betweenthe other lines and the recurrent parent. The 1000-grain weight for line230081 was 21.5 grams, which was the smallest among the lines, and sig-nificantly smaller than that of Cypress (25.1 grams/1000 grains–1). Therewere no significant differences between the other lines and the recurrentparent for 1000-grain weight, although grains of some of them werelighter or heavier than those of the recurrent parent (Table 4).

The ratio of the grain length to width for Cypress was 3.7, a typicallong grain. The ratio of the grain length to the width for line 230413 was2.9. Therefore, this line would not be considered a typical long-graintype (USA Rice™ Federation, 2002). The ratios of grain length to widthfor the other lines ranged from 3.44 to 3.84, which being more than 3.0would be classified as typical long-grains (Table 4).

There were no significant differences in percentages of brown rice andwhole (head) rice, between the backcross lines and Cypress (Table 5), buttotal milled rice for line 230221 was 64.11%, which was significantly

58 JOURNAL OF CROP IMPROVEMENT

TABLE 4. Comparison of the major grain characteristics between Cypressand lines developed from repeated backcrosses using Cypress as the Recur-rent parent. Based on grain samples collected from the 2001 and 2002 yieldtests combined.†

Entries Mean GrainLength (mm)

Mean GrainWidth (mm)

Ratio of GrainLength to Width

Mean Grain Weight(grams/1000 grains)

Cypress 9.37a 2.53c 3.70 25.07ab230081 8.70c 2.53c 3.44 21.47c230221 9.27ab 2.60bc 3.56 24.80b230251 9.33ab 2.70b 3.45 25.87ab230281 9.03b 2.53c 3.57 24.93b230408 9.07b 2.63bc 3.45 25.07ab230413 8.80c 3.03a 2.90 26.13a230420 9.60a 2.50c 3.84 24.80b†Means followed by different letters are significantly different at p = 0.05.

Dow

nloa

ded

by [

Dr

Man

jit S

. Kan

g] a

t 03:

42 0

9 A

ugus

t 201

5

lower than that of Cypress (71.41%). The rest of the lines were not sig-nificantly different in total milled rice than in Cypress (Table 5).

All the seven selected lines were resistant to Liberty. This indicatedthat the Bar gene was successfully transferred into these lines throughrepeated backcrossing. The data also indicated that the phenotypes of theseven selected lines from repeated backcrosses, when Cypress was usedas the recurrent parent, were recovered.

Yield Tests Comparing Five Backcross Lines Developedwith Lafitte as the Recurrent Parent

Lafitte is a medium grain cultivar, which is similar in grain shape to theBar gene donor parents, Taipei 309 and Nipponbare. The mean yieldsfrom the selected backcross lines varied from 5923.6 kg ha–1 to 7210.9kg ha–1 (Table 6). The yield of line 230643 was 7210.9 kg ha–1, whichrepresented the highest and significantly higher yield than Lafitte (6158.8kg ha–1). There was no significant difference in yields between Lafitteand the other tested lines (Table 6). There were significant differencesin plant height among the tested lines. Plant height of experimental line230643 was 96.6 cm, which was significantly greater than that of theother lines and Lafitte. The days-to-heading varied considerably. Ex-perimental line 230629 had 79.2 days from planting to heading, whichwas significantly smaller than that of the rest of the backcross lines. Ex-perimental lines 230643 and 231809 had about the same heading time,and lines 230485 and 230706, and Lafitte had about the same time-to-heading (Table 6).

Shao et al. 59

TABLE 5. Comparison of the milling characteristics between Cypress andbackcross lines, developed from repeated backcrosses using Cypress as therecurrent parent, in the 2001 and 2002 yield tests.*

Entries Mean Brown Rice (%) Mean Milled Rice (%) Mean Whole Rice (%)

Cypress 79.14a 71.41a 56.91a230081 78.69a 66.11ab 56.88a230221 77.85a 64.11b 56.55a230251 77.83a 64.96ab 57.45a230281 78.53a 65.44ab 57.06a230408 79.29a 64.79ab 56.32a230413 78.25a 68.33ab 55.96a230420 78.48a 65.94ab 58.46a

*Means not followed by the same letter were significantly different at p = 0.05.

Dow

nloa

ded

by [

Dr

Man

jit S

. Kan

g] a

t 03:

42 0

9 A

ugus

t 201

5

There was no significant difference in grain length between Lafitteand the backcross lines developed using Lafitte as the recurrent parent.Also, there was no significant difference in grain width between all thelines and Lafitte. The 1000-grain weight of lines 230706, 230485,230643, and 231809 was significantly lighter than that of Lafitte (24.9grams 1000-grains–1). No significant difference in 1000-grain weightexisted between Lafitte and line 230629 (Table 7). The ratios of grainlength to width for all the tested entries were between 2.0 and 2.9 (Table 7).All the lines developed using Lafitte as the recurrent parent were typicalmedium grain types.

There was no significant difference in percentages of brown rice ortotal milled rice, between Lafitte and the backcross lines (Table 8). Thewhole (head) rice of line 230706 was 52.9%, which was significantlylower than that of Lafitte (59.7%). The percentage whole rice for theother backcross lines was not significantly different from that of Lafitte(Table 8).

The data from field progeny rows and two years of yield testing indi-cated that all five of these selected lines were resistant to Liberty, sug-gesting that these lines had the Bar gene. Liberty resistance was laterconfirmed by PCR. It appears that repeated backcrosses using Lafitte asthe recurrent parent recovered the major agronomic characteristics ofLafitte along with the addition of Liberty resistance.

Existence of Bar Genes in the Elite Lines from Backcrosses

Repeated application of Liberty in field selection tests, conductedfrom 1997 to 2002 and in yield tests conducted in 2001 and 2002 with

60 JOURNAL OF CROP IMPROVEMENT

TABLE 6. Comparison of yield and agronomic characteristics between Lafitteand five backcross lines selected from repeated backcrosses using Lafitte asthe recurrent parent in the 2001 and 2002.*

Entries Mean Yield(kg/ha)*

Mean PlantHeight (cm)

Mean Days-to-Heading (days)

Lafitte 6158.8b 89.8b 83.8a230485 6046.4b 89.5b 84.0a230629 6359.9ab 88.2b 79.2c230643 7210.9a 96.6a 81.1b230706 5923.6b 82.7b 83.8a231809 6020.5b 87.0b 81.5b

*Mean values not followed by the same letter were significantly different at p = 0.05.

Dow

nloa

ded

by [

Dr

Man

jit S

. Kan

g] a

t 03:

42 0

9 A

ugus

t 201

5

Liberty applications, showed that the backcross lines were resistant toLiberty.

PCR evaluation of leaf blade samples from the Liberty-resistantentries from the yield tests indicated that the Bar gene was present in theexperimental lines. No gene amplification band was identified for anyof the non-transgenic recurrent parents, that is, Cypress and Lafitte. PCRreactions for lines 230081, 230221, 230225, 230281, 230408, 230413,and 230420 from the repeated backcrosses using Cypress as the recurrentparent, and lines 230485, 230629, 230643, 230706, and 231809 devel-oped using Lafitte as the recurrent parent, were all positive for the Bargene. Also, the two Bar gene donor parents, N-84 and T-64, were posi-tive. The Bar gene in these lines could be amplified and expressed throughPCR (Figures 1 and 2).

Shao et al. 61

TABLE 8. Comparison of the milling data between Lafitte and the backcrosslines developed through repeated crosses using Lafitte as the recurrent parentin the 2001 and 2002 yield tests.*

Entries Mean Brown Rice (%) Mean Milled Rice (%) Mean Whole Rice (%)

Lafitte 77.69a 66.67ab 59.66a230485 76.98a 66.25ab 57.75ab230629 78.60a 71.38a 58.20a230643 77.89a 66.76ab 56.12ab230706 77.37a 62.63b 52.91b231809 77.75a 69.76a 57.34ab

*Mean values not followed by the same letters were significantly different at p = 0.05.

TABLE 7. Comparison of grain characteristics between the cultivar Lafitte andbackcross lines selected from repeated backcrosses using Lafitte as the recur-rent parent in 2001 and 2002 yield tests.*

Entries Mean GrainLength (mm)

Mean GrainWidth (mm)

Ratio of GrainLength to Width

Grain Weight(g/1000 grains)

Lafitte 7.90a 3.03a 2.34 24.93a230485 7.73a 3.00a 2.58 23.79b230629 7.80a 3.00a 2.60 24.80a230643 8.07a 2.77a 2.91 23.47b230706 7.87a 3.00a 2.62 22.93b231809 7.67a 3.00a 2.56 23.60b

*Mean values not followed by the same letters were significantly different at p = 0.05.

Dow

nloa

ded

by [

Dr

Man

jit S

. Kan

g] a

t 03:

42 0

9 A

ugus

t 201

5

DISCUSSION

The general principle for the use of repeated backcrosses is that therecurrent parent is not improved except for the character of interesttransferred from the donor parent. It has been demonstrated that re-peated backcrossing is an effective way to transfer the target gene orcharacter to the recurrent parent (Shao et al., 1991; Yan et al., 1991).The character to be transferred from the donor parent must have simpleinheritance. If a character is controlled by a single dominant gene, itis relatively easy to transfer the gene into the recurrent parent. For traitscontrolled by polygenes, however, this method is not the most suitable one.

Three transgenic lines, T-28 and T-64 from Taipei 309, and N-84 fromNipponbare, were used as sources of resistance to Liberty resistance andthus, as donor parents. The commercial rice cultivars, Cypress and

62 JOURNAL OF CROP IMPROVEMENT

1

Kb

1.6

1.0

0.8

0.60.5

0.2

2 3 4 5 6 7 8 9 10 11 12

FIGURE 1. PCR analysis for Bar genes in lines from repeated backcrosses.Lane 1: 1 Kb molecular marker; lane 2: non-transgenic recurrent parent Cy-press; lanes 3 to 9 represent experimental lines 230081, 2300221, 230281,230251, 230408, 230413, and 230420; lanes 10 and 11 refer to their transgen-ic parents N-84 and T-64. Lane 12 refers to the positive control from pBSL 21.

Dow

nloa

ded

by [

Dr

Man

jit S

. Kan

g] a

t 03:

42 0

9 A

ugus

t 201

5

Lafitte, were used as the recurrent parents. The results from progenyrow selections and two years of field-testing indicated that the primaryphenotypes of the recurrent parents could be recovered after three tofive repeated backcrosses, and the Liberty resistance gene was trans-ferred into the selected lines. Comparison of the data collected on themajor agronomic characteristics of lines generated from different gen-erations of repeated backcrosses in the field test of 2000 indicated thatphenotypes of the selected lines changed gradually, and generally thephenotypes of lines from BC3 were very similar to the recurrent parents(Table 2). This was also confirmed via the two-year yield tests. In BC3,93.75% of the genotype of the recurrent parent is theoretically recover-ed. When Cypress was used as the recurrent parent, all seven selectedlines were resistant to Liberty, and lines 230221 and 230251 had verysimilar phenotypes to Cypress for yield, plant height, days-to-heading,grain length, grain width, 1000-grain weight, total brown rice, totalmilled rice, and whole rice (Tables 3-5). When Lafitte was the recurrent

Shao et al. 63

1

kb

1.6

1.0

0.8

0.60.5

0.2

2 3 4 5 6 7 8

FIGURE 2. PCR analysis of lines selected from repeated backcrosses andnon-transgenic cultivars used as recurrent parents in the backcross program.Lane 1: molecular marker; lane 2 is negative control; lanes 3 to 5 representlines 230485, 230629, and 230643; lane 6 refers to non-transgenic Lafitte;lanes 7 to 8 refer to lines 2300706 and 231809.

Dow

nloa

ded

by [

Dr

Man

jit S

. Kan

g] a

t 03:

42 0

9 A

ugus

t 201

5

parent, the five selected backcross lines, 230485, 230629, 230643,230706, and 231809, were resistant to Liberty, and their phenotypes inall the investigated characteristics were close to that of the recurrentparent (Tables 6-8). When a “similar phenotype” was indicated, itmeant that there were no or few significant differences between each ofthe lines and the recurrent parents for the characteristics measured.

The genetic background of the recurrent parent is not usually consid-ered when the repeated backcross method is used for gene transfer fromthe donor parent to the recurrent parent. In actuality, the genetic back-ground may have major effects on the phenotypes of the lines selectedfrom repeated backcrosses. This method using the same donor parent,but different recurrent parents, will produce different results. In otherwords, the efficacies for the backcross line selection are different. Theresults from the two years of comparative experimental tests showedclearly that the ease of selection of backcross lines with Liberty resis-tance using different cultivars as the recurrent parents varied.

Backcross line selection when Cypress and Lafitte were used as therecurrent parents was easier than when the other cultivars were used asthe recurrent parents. The experimental lines 230251 and 230281,which came from repeated backcrossing using Cypress as the recurrentparent, had phenotypes very similar to the phenotypic characteristics ofCypress. The phenotype of line 230643 was very similar to the pheno-type of its recurrent parent, Lafitte. However, when Lemont, Katy, andDrew were used as the recurrent parents, it was more difficult to selectbackcrossed lines with the similar phenotype of the recurrent parent(data not shown).

The yields from two years of yield tests conducted in 2001 and 2002showed that plots not sprayed with Liberty yielded significantly higherthan plots sprayed with Liberty, even though the plants were resistant toLiberty. Leaf discoloration was also observed after Liberty applicationon the leaves of resistant plants, but the plants did not die. The resultsfrom greenhouse tests showed that Liberty’s inactive ingredient solu-tion also caused leaf discoloration on both non-transgenic and transgen-ic plants. The Liberty herbicide formulation caused the same leafdiscoloration when the same concentration was applied to transgenicplants. It is possible that inactive ingredients in the formulation causedsome of the loss following application of Liberty. Damage from appli-cation of Liberty to transgenic rice may by reduced or eliminated bymodifying Liberty formulation.

64 JOURNAL OF CROP IMPROVEMENT

Dow

nloa

ded

by [

Dr

Man

jit S

. Kan

g] a

t 03:

42 0

9 A

ugus

t 201

5

REFERENCES

Allard, R.W. 1960. Principle of plant breeding. John Wiley & Sons, Inc. New York.Buerstmayr, H., M. Lemmens, G. Fedak, and Ruckenbauer. 1999. Back-cross recipro-

cal monosomic analysis of Fusarium head blight resistance in wheat (Triticumaestivum L.) Theor. Appl. Genet. 98:76-85.

Christou, P. 1997. Rice transformation: bombardment. Plant Mol. Biol. 35:197-203.Fehr, W.R. 1987. Principles of cultivar development. Volume 1: Theory and technique.

Macmillan Publishing Company, New York.Nelson, R.R. 1973. Breeding plants for disease resistance: concepts and application.

The Pennsylvania State University Press. Pennsylvania.Rush, M.C., X.Y. Sha, Q.M. Shao, S.D. Linscombe, and D.E. Groth. 1999. Develop-

ment of sheath blight resistance in rice. Pp. 56-79. 91st Annual Research Report–Rice Research Station, Louisiana Agricultural Experiment Station, Louisiana StateUniversity Agricultural Center, Crowley, LA.

Sambrook, J., E.F. Fritsch, and T. Maniatis. 1989. Molecular cloning–a laboratorymanual. Second edition. Cold Spring Harbor Laboratory Press.

SAS Institute. 2000. SAS guide for personal computers. 1st ed. SAS Inst., Cary, NC.Shao, Q.M., and L. Zhu 1991. Continuous backcrosses as related to the inheritance of

bacterial blight. J. Southwest Agri. Univ. 13:292-295.Shao, Q.M., L. Zhu, and J.M. Wan. 1992. Genetics studies on resistant to bacterial

blight in eight breeding lines of rice. J. Southwest Agri. Univ. 14:35-39.USA Rice™ Federation. 2002. The USA Rice Council. Available online at http://

www.riceprocessing.com/about_fed_council.htm.Wan, J.M., L. Zhu, and Q.M. Shao. 1991. Breeding of derived resistant lines in Indica

rice and their genetic analysis of resistance to bacterial blight. J. Nanjing Agri.Univ. 13:1-18.

Xie, Q.J., M.C. Rush, J. Oard, and S.D. Linscombe. 1995. Development of an efficientBiolistic PDS-1000/He mediated transformation system and regeneration of riceplants with a high level of ignite herbicide resistance. Pp. 112-115. 87th Annual Re-search Report–Rice Research Station, Louisiana Agricultural Experiment Station,Louisiana State University Agricultural Center, Crowley, LA.

Yan, L. P. 1991. Development of “three lines” hybrid rice in China. Sinica AgriculturalResearch Academy Press, Beijing, China. Pages 11-50.

Zemetra R.S., J. Hansen, and C. Mallory-Smith. 1999. Potential for transfer of herbi-cide resistance via wheat � jointed goatgrass hybrid. Available online at http://www.css.orst.

Shao et al. 65

Dow

nloa

ded

by [

Dr

Man

jit S

. Kan

g] a

t 03:

42 0

9 A

ugus

t 201

5