DEVELOPMENT OF EARLY FLOWERING, CANOLA-GRADE BRASSICA JUNCEA GERMPLASM

Rex N Oram¹, Philip A Salisbury², John T O Kirk¹, and Wayne A Burton³

¹CSIRO Plant Industry, G.P.O. Box 1600, Canberra A.C.T., Australia 2601

²Faculty of Agriculture, University of Melbourne, Parkeville, Vic., Australia 3052

³Victorian Institute for Dryland Agriculture, PMB 260, Horsham Vic., Australia 3400

ABSTRACT

 This paper reviews the modernisation of Brassica juncea by reducing the erucic acid content of the oil to less than 1% ( Kirk and Oram 1981) and the glucosinolate content of the oil-free meal to less than 30 mmole/g, so that it can be fed as a high protein supplement to monogastric and ruminant animals.  Backcrossing the low erucic acid genes into the Canadian condiment cultivar Domo gave cv. Siromo, which enabled the establishment of the cold-pressed mustard seed oil industry by the Yandilla Mustard Oil Enterprise at Wallendbeen, NSW in 1989.  Seed glucosinolates then were reduced in Australia by somaclonal variation, by reconstituting B. juncea from a low glucosinolate B. rapa cultivar, by repeated mutagenesis with EMS and g-irradiation, and by recombining all of these sources of variation.  Concurrently, crossing between the original long-day requiring, low erucic acid stocks and early flowering, day neutral accessions from India and southern China, followed by selection in the NSW and Victorian wheatbelts and intercrossing, gave early maturing, low erucic acid lines which yielded 15-20% more than the standard Indian accession.  Then a low glucosinolate genotype developed in Canada was crossed to CSIRO and VIDA selections to give early-flowering, high-yielding double low B. juncea lines. Recently, the oleic acid concentration in the oil has been increased to 54% by reducing linolenic acid to 7%. The performance of the current double low selections in dry areas will be described by Burton et al in these Proceedings, as are the additional changes required to make “juncea canola” interchangeable with “napus and rapa canola.”

KEYWORDS: seed composition, erucic acid, glucosinolates, double low, high yield.

INTRODUCTION

Indian mustard (Brassica juncea (L.) Czern. & Coss.) has several potential advantages over canola (B. napus L.) as an oilseed in semi-arid regions.  In the southern Australian wheatbelt, these regions occupy three million hectares (JF Angus, personal communication) and receive 225-350 mm of rain per annum, mainly in the cooler half of the year.  No oilseed crop is well adapted to these regions, but one is needed for inclusion in cereal/pasture rotations to control the weeds, pests and root diseases which become prevalent during the cereal phase.  Therefore, a research program has been conducted for the past 25 years with the objective of creating B. juncea genotypes which produce seeds that are interchangeable with those of B. napus canola.  The principal advantages of B. juncea over B. napus are as follows.  B. juncea seeds will germinate in soil which is too dry for canola germination (H-H Muendel, personal communication).  B. juncea seedlings are more vigorous, and hence cover the soil surface more quickly, thus using more of the early rainfall for transpiration and growth and wasting less by evaporation from the soil surface.  B. juncea can produce twice as much above-ground biomass and seed as canola under controlled drought conditions (Wright et al 1995).  Relative to B. napus, B. juncea also is more tolerant to the blackleg fungus (Leptosphaeria maculans), its seed pods shatter less readily, and its seeds potentially have a higher percentage of oil + protein because the yellow seed coat is thinner. Both species have oils low in saturated fatty acids.

This paper describes improvements in seed quality in Australia through reductions in glucosinolate and linolenic acid concentrations, and increases in oleic acid content, as well as progress towards achieving higher yield in earlier maturing, double low recombinants in B. juncea.

CREATING LOW GLUCOSINOLATE B. JUNCEA

The various steps taken in Australia to reduce the concentration of glucosinolates in the seeds are shown in Figure 1. In 1982, seeds of the original low erucic acid accession, Zem 1, were irradiated with g-rays and seeds of M₂ plants were ‘hammer-tested’ (Lein 1970) to detect the reduced glucosinolate mutants LG 1-3 and LG 4-3. Secondly, B. juncea was re-synthesised from B. nigra and low glucosinolate B. rapa. The product was self-incompatible, so it was crossed to B. juncea cv. Lethbridge to produce Syn Y. Thirdly, a reduced glucosinolate somaclonal variant was obtained in the State Chemistry Laboratory of Agriculture Victoria by tissue culture of an Indian accession (Palmer et al 1988). This variant was recombined with an M₄ selection from g-irradiated Syn Y, and an F₄ selection was crossed with an F₃ plant from Syn Y ´ LG4-3 to produce the first unstable double low population, Family 167 (Figure 1; Table 1; Oram and Kirk 1993).  Mutants with very low glucosinolate concentrations were induced in Family 167 by treatment of successive generations with sodium azide, ethyl methanesulphonate and g-rays.  Seeds from individuals in the next generation (M_2,3,4) were tested firstly for glucosinolates by the ‘hammer test’ and secondly for reduced linolenic acid by the thiobarbituric acid method of McGregor (1974). Suspected low glucosinolate or low linolenic acid families were analysed respectively by X-ray fluorescence spectroscopy (Pinkerton et al 1993) or by gas-liquid chromatography.  Two low glucosinolate mutants were induced, including ‘Apps 1534’ (Figure 1).  Family 167 also was hybridised with a Canadian reduced glucosinolate, high erucic acid, partially sterile, late-flowering selection (Agriculture Canada number DBBS 18-008) having brown, shrivelled seeds, which was derived by crossing a very low glucosinolate B. rapa with a B. juncea containing predominantly 3-butenyl glucosinolate in its seeds, then backcrossing the F₁ to the same B. juncea cultivar (Love et al 1990).  The cross gave late-maturing segregants in family 651 with lower glucosinolate levels than either parent.  The descendants of Apps 1534 and family 651 were crossed to 30 accessions from six countries to broaden the genetic base of the double low population, and F₂ selections were backcrossed to Apps 1534 and family 651 derivatives.  Further selection gave families such as 766-5-1 and 887-2-6 which produced seed with less than 5 mmol glucosinolates/g seed (Table 1).  DBBS 18-008 also was crossed to an early-flowering, high-yielding breeding line to produce the double low family, RP116, which, though late-flowering and short-podded, also contributed significantly to the current advanced material (Figure 1).

Table 1.  Glucosinolate concentration (mmol/g oil-free meal) in some original, high

                glucosinolate stocks and in successive reduced and low glucosinolate breeding lines.

Line	Type	Origin	Glucosinolate conc.	Flowering time^1
PI 183117	Accession	India	182	Early
Syn Y	Re-synthesised B. juncea		176	Late
LG 12-2	g-ray mutant	Syn Y	156	Late
SCL 4	Somaclonal variant	PI 183117	144	Mid-season
167-2-5-21	ü   Recombined		74	Late
651-2-5-7	î   mutants and		19.2	Late
815-1-6-1	ì   interspecific		7.2	Mid-late
887-1-1	ï    hybrids		4.9	Early
766-5-1	þ		3.5	Early-mid

¹ When sown in autumn at latitude 35°S.

                                                   CPI       651

   CPI                                       107520                                   688

 81796                                        ^China  324 892

^India 99Y-1-1                                   651

 Zem 1

  ^China                CPI                                                    347

                     61690

    CPI              ^India                                                                                                      875

  61690                                       292-95  923

^India 397 651

Zem 1      ^g-rays            Early Zem-5

 ^China

                CPI 81792         CPI 96866

^{India                             China}

   Zem 1 292-62 RP 116-130

^China

                                                                                        76883                                651

CPI 81792                                                                          ^China               373 887

^India DBBS 18-008 292-62                                 693  

 ^Canada

RP 116-139

                              PI 347615                                                                            675

 ^India

96-5

LG 1-3-10

Zem-1    ^g-rays 766     JM27 - 30

^China DBBS 18-008

                             LG 4-3                                                                     ^Canada

651

B. nigra

^Yugoslavia 733

B. rapa                         Syn Y                                     167 815

^{cv. Candle} g-rays                                                    g-rays, EMS, azide 693

         B. juncea                          LG 12-2                                        Apps 1534

               ^{cv. Lethbridge}

    PI 183117       ^Somacloning       SCL 4                                                                          731       

^India    91-23896 (Canada)

JM01 RP 116-139 750

                                                    1-193                                                                             652

         CPI 61687

CPI 61694                      JM02-25 ^India 880     JM31-33

^India                                                                        RP 116-130

Zem 2

                                           cv. Varuna                            JM26

^India

                                                       Zem 1

Figure 1. Pedigrees of some Australian low erucic acid and double-low breeding lines of Brassica juncea.  The countries of origin of the parents are shown below the accession numbers: CPI = Commonwealth (of Australia) Plant Introduction, PI = (US) Plant Introduction. Others are Australian breeders’ code names and numbers, except DBBS 18-008 and 91-23896, which are Canadian reduced glucosinolate lines segregating for erucic acid levels. Bold type shows low erucic acid lines, italics show low glucosinolate lines.  Double low lines used frequently in crosses are shown in rectangular or elliptical boxes.  The advanced lines on the right have been (1) re-selected to give the current JM lines and (2) inter-crossed to generate a wider gene-pool for later selection.

YIELD IMPROVEMENT IN DOUBLE LOW B. JUNCEA

The low erucic acid trait was discovered in long day-requiring condiment lines from northern China (Kirk and Oram 1981).  This China-East European type flowers late and grows to a height of 3 m when sown in autumn in south-eastern Australia.  Day-neutral accessions from south Asia are much better adapted, flowering three weeks earlier and growing to heights of 1.5-2 m.  Two rounds of crossing and selection for higher yield between Indian or Chinese accessions and the low erucic donors, Zem 1 and Zem 2, gave a 15-20% increase in yield (e.g., family 397, Figure 1, Table 2). 

The low erucic acid selections have contributed to the recent, dramatic increases in the yields of double low B. juncea breeding lines, mainly by conferring early maturity (Table 2).  However, double low B. napus cultivars were more productive than B. juncea even at the driest sites.  Napus canola had a much shorter flowering period than B. juncea; even though the highest yielding napus canolas began flowering seven days later than the earliest B. juncea, they finished flowering twelve days earlier (Oram et al 1997).  Therefore, a primary selection goal in double low mustard has been to reduce the period from sowing to the end of flowering by at least two weeks. 

Table 2.  Yield of original and advanced lines of B. juncea at dry interstate trial sites

					Mean yield (t/ha)
Line	Type	Flowering	1995 5 sites	% of 81792		1996 7 sites	% of 81792	1997 8 sites	% of 81792
81792	Indian accession	Early	0.74	-		0.89	-	1.13	-
397-23-2-	Low erucic ü	Early	0.86	116		1.07	110	1.37	121
651-2-5-7	Double low î Breeding	Late	0.57	77		-	-	-	-
766-1-6-1	Double low ì lines^1	Early-mid	-	-		0.56	63	0.84	74
887-2-6-1	Low gluc.   þ	Early	-	-		-	-	1.46	129
Hyola 42	Hybrid B. napus canola	Early	-	-		1.18	132	1.59	141

¹ Pedigrees of the breeding lines are shown in Figure 1; ‘Low gluc’= low glucosinolate, segregating for erucic acid.

Two other yield-limiting defects, partial sterility and shrivelled seeds, were generated by the extensive use of mutagens, somacloning and interspecific hybridisation during the creation of the low glucosinolate type.  Both defects are being corrected by selection in the broader, low glucosinolate gene pool.  More fertile, plump-seeded, earlier selections have yielded more seed than napus canola at dry Victorian sites in 1998 (Burton et al, these Proceedings).  Unexpectedly, some Australian material proved to be earlier maturing than B. rapa and higher yielding than Canadian B. juncea and B. napus lines in the first year of trials in Canada in 1998 (WA Burton, personal communication).

FATTY ACID MODIFICATION

The standard level of 15% linolenic acid in low erucic acid mustard oil is acceptable in a cold-pressed, cooking and salad oil such as Yandilla Mustard Seed Oil (Lehane 1996), but a reduction to 9-11%  is  needed  to  match  canola oil.   Variants  with  this level  have been  found in  the  Chinese

Table 3.  Fatty acid composition of a low erucic Indian mustard cultivar, Siromo, and of

                modified breeding lines.

Line	Palmitic	Stearic	Oleic	Linoleic	Linolenic	Erucic	Other C20-C24
Siromo	3.7	2.1	44.7	31.4	14.9	0.30	2.7
1006-4	4.0	3.5	49.2	33.0	6.5	0.73	3.1
1043-27	3.3	4.8	54.3	28.4	7.3	0.00	1.9

accessions which contained the low erucic acid alleles (J.T.O. Kirk and C.J. Hurlstone, unpublished), and also in mutagenised populations (Oram and Kirk 1993).  Crosses between these and improved breeding lines gave an F₂ segregate, 1006-4, with concentrations as low as 6.5% (Table 3).

No mustard genotypes with oleic acid levels above 55% have yet been found in the CSIRO germplasm collections or in three mutagenised populations screened for lower linolenic acid levels.  A concentration of at least 65% is needed to match the oleic acid level in napus canola.  However, one reduced linolenic acid segregate recently was found with 54.3% oleic acid (1043-27 in Table 3).  Further reductions in linoleic, linolenic and C20-C24 fatty acids seem possible: this should lead to further increases in oleic acid concentrations. Recently, Potts et al (these Proceedings) discovered a high oleic variant in B. juncea, and Stoutjesdijk et al (these Proceedings) have down-regulated D-12 desaturase activity by a molecular mechanism, leading to oleic acid levels as high as 73% in hemizygotes.  Future collaborative work to increase oleic acid levels in juncea canola will be discussed by Burton et al in these proceedings.

ACKNOWLEDGMENTS

B. juncea R & D has been financially supported in recent years by the Australian Grain Research and Development Corporation.  The collaboration and assistance of CJ Hurlstone, DJ Robson, T Golebiowski, JP Edlington, SJ Pymer, CF Greenwood, PE Veness, the late ID Wiencke, B Kay and G Bell is gratefully acknowledged.

REFERENCES

Kirk JTO, and Oram RN  1981  Isolation of erucic acid-free lines of Brassica juncea: Indian    mustard now a potential oilseed crop in Australia. Journal of the Australian Institute of Agricultural Science.  47: 51-2

Lehane R  1996  Towards a mustard alternative. Rural Research, CSIRO Australia. 171: 16-20.

Lein KA  1970  Quantitative Bestimmungsmethoden für Samenglucosinolate in Brassica-arten   und ihre Anwendung in der Züchtung von glucosinolatearmen Raps.  Zeitschrift für Pflanzenzüchtung 63: 137-54.

Love HK, Rakow G, Raney JP, and Downey RK  1990  Development of low glucosinolate      mustard.  Canadian Journal of Plant Science 70: 419-24.

McGregor DI  1974  A rapid and sensitive spot test for linolenic acid levels in rapeseed. Canadian Journal of Plant Science 54: 211-3.

Oram RN, and Kirk JTO  1993  Induction of mutations for higher seed quality in Indian mustard. In:      Proceedings of the 10th Australian Plant Breeding Conference, Gold Coast, Qld. pp. 187-91.

Oram RN, Walton G, Marcroft S, Potter TD, Burton WA, Castleman GH, Easton AA, and      Kirk JTO  1997  Progress in breeding canola-grade Indian mustard.  In: Proc. 11th            Aust. Res. Assembly on Brassicas. Agriculture WA: Perth. pp 79-83.

Palmer MV, Sang JP, Oram RN, Tran DA, and Salisbury PA  1988  Variation in seed glucosinolate concentrations of Indian mustard (Brassica juncea (L.) Czern. & Coss.) Australian Journal of Experimental Agriculture 28: 779-82.

Pinkerton A, Randall PJ, Wallace PA, Vonarx MM, and Mailer RJ  1993  Determination of      total glucosinolates in oilseed rape by X-ray spectrometric analysis for oxidised sulphur           (S⁶⁺). Journal of the  Science of Food and Agriculture 61: 79-86.

Wright PR, Morgan JM, Jessop RS, and Cass A  1995  Comparative adaptation of canola (Brassica    napus) and Indian mustard (B. juncea) to soil water deficits: yield and yield           components.  Field Crops Research 42: 1-13.