ADAPTATION OF CANOLA TO A MEDITERRANEAN ENVIRONMENT: GENETICS OF THE VERNALIZATION RESPONSE

S.R. DAHANAYAKE* and N.W. GALWEY

*Division of Botany, Faculty of Natural Sciences, The Open University of Sri Lanka, P.O. Box 21, Nawala, Nugegoda, Sri Lanka

Department of Plant Sciences, Faculty of Agriculture, The University of Western Australia, Nedlands, WA 6907,  Australia.   Email: ngalwey@cyllene.uwa.edu.au

ABSTRACT

The capacity for vernalization varies among genotypes of spring rape, Brassica napus var. annua, and affects their suitability for cultivation in the mediterranean environment of Western Australia. The inheritance of the vernalization response was investigated in crosses among five fully inbred genotypes of spring rape derived from a cross between two cultivars differing in their vernalization response. Generation mean analysis of time to flowerig and other related characters indicated that additive, dominance and non-allelic interaction effects all play an important role in the vernalization responses of these characters. Estimates of broad- and narrow-sense heritability indicated that the vernalization responses can be fairly readily eliminated by selection. The potential of a combined molecular and cytogenic approach to the mapping of loci that confer a vernalization response was also explored in backcrosses of B. napus (genomic constitution AACC) to B. rapa (AA). The RFLP probe/enzyme combination wg7b3/HindIII co-segregated with lateness at a locus on the C genome. Screening of a larger number of backcross progeny would permit the identification of the C-genome chromosome conferring the vernalization response, which would facilitate marker assisted selection to eliminate the vernalization requirement for early flowering.

KEYWORDS: Brassica napus var. annua, Brassica rapa, generation mean analysis, heritability, restriction fragment length polymorphism (RFLP), cytogenetics

INTRODUCTION

Until recently all rapeseed cultivars grown in Western Australia since the establishment of the industry in 1969 were introduced directly from various sources in the Northern Hemisphere. These were spring-sown types (Brassica napus var. annua L.) from regions with a summer growing season, and it is perhaps not surprising that they show substantial variation in developmental pattern in this region (Thurling and Vijendra Das, 1977), which has a mediterranean climate with a wet winter growing season and a dry summer. In particular, it was to be expected that the capacity for vernalization (promotion of flowering by low temperature) would vary among these genotypes.   The present investigation aimed to determine the overall genetic control of this vernalization response, including possible polygenic effects in addition to the major genes already detected (Thurling and Vijendra Das, 1979a, b).   In addition, an attempt was made to identify marker loci associated with the vernalization response and to locate them on a particular chromosome, in order to increase their value for marker-associated selection.   Such markers would provide a non-destructive method for early recognition of progeny with non-vernalization responsive genotypes in a breeding program.   For this purpose, a vernalization-responsive genotype of B. napus was backcrossed to a non-responsive genotype of Brassica rapa, and phenotypic, molecular and cytological observations were made on the progeny.

MATERIALS AND METHODS

Quantitative analyses

Five genotypes of Brassica napus var. annua, TB 14, TB 16, TB 28, TB 29 and TB 33, derived from a cross between the cultivars Target (vernalization insensitive) and Bronowski (vernalization sensitive – Thurling and Vijendra Das, 1977) were chosen for the diversity of their vernalization responses. TB 14, which showed a strong vernalization response with regard to time to flowering and a moderate vernalization response with regard to number of leaf nodes at flowering and stem height at flowering, was crossed with the other genotypes. The F₁ progeny were selfed to generate F₂ populations and backcrossed to the respective parents (P₁ and P₂) to produce two first backcross progenies (BC₁ and BC₂) for each of the four crosses.   The seeds of the six populations (P₁, P₂, F₁, F₂, BC₁ and BC₂) for each cross were exposed to low temperature (4°C) in the dark on moist filter papers wetted with 1200 Kpa (302.44 g/l) polyethylene glycol solution for 8 weeks. In each case another batch of seed was left untreated. Seeds were then allowed to germinate on moist filter papers and the seedlings were grown in pots, 13 cm in diameter, one seedling per pot. The standard commercial potting mixture was used with the fertilizer Osmocote.

Generation means were obtained from plants grown in a controlled environment room in which the main light period of 10h was provided by Na-vapour and Metal Halide lamps giving a radiation level of 168 mE m^-2 s^-1 at the pot surface. Saturating long photoperiod (18h) was obtained with a 75W incandescent lamp (7.5 mE m^-2 s^-1). The post-vernalization temperature was maintained at 15°C. There were 10 plants of each population and the pots were arranged randomly. The date on which the first open flower appeared (time to flowering), the number of leaf nodes on the main stem at flowering and the length of the main stem from cotyledons to shoot apex at flowering were recorded for each plant.   Variance components and heritability were studied in two crosses, namely TB 14 ´ TB 28, in which the parents had widely different vernalization responses and TB 14 ´ TB 33, in which the vernalization responses of the parents were fairly similar. The post-vernalization conditions were similar to those in the generation mean experiment. There were 20 plants of each population and the pots were arranged randomly.

Molecular analyses

A combination of cytogenetic and molecular genetic techniques can be used to aid the genetic mapping of the vernalization response in B. napus. For this purpose, it is assumed that the vernalization response in B. napus var. annua is determined by recessive alleles at each of two loci, one (v₁) on a chromosome of the A genome and the other (v₂) on a chromosome of the C genome.   The backcross of vernalization-responsive B. napus to non-responsive B. rapa then produces the following genotypes:

B. napus (10AA – v₁v₁) + (9CC – v₂v₂)    ´ B. rapa (10AA – V₁V₁)

F₁ hybrid:  (10AA – V₁v₁) + (9C – v₂)

BC generations: (10AA – V₁v₁ or V₁V₁) + a variable number of C chromosomes, sometimes carrying v₂.

The genotype TB 14 of B. napus var. annua, which is vernalization-responsive for flowering, was crossed as the female parent with the non-vernalization responsive B. rapa cv. IB-5 (an annual). F₁ plants were used as females in crosses with B. rapa  to produce 35 BC₁ plants.   Six late-flowering BC₁ plants were used to produce 16 BC₂ plants. The time from sowing to the appearance of the first open flower was recorded in the parental, F₁, BC₁ and BC₂ generations.   DNA was extracted, digested with enzymes EcoRI, EcoRV, DraI and HindIII and probed with DNA clones wg5a5, wg6b10, wg7b3, wg7f3 and wg8g1 (provided by Prof. T. C. Osborn, University of Wisconsin, USA). These probes were isolated from a PstI genomic DNA library from  B. napus cv. ‘Westar’. The number of chromosomes in somatic cells of BC₁ and BC₂ plants was determined from root tip squashes.

RESULTS

Quantitative analyses

Scaling tests (Mather and Jinks, 1971) gave values significantly different from zero for at least one of the three characters measured on both the low-temperature treated and the control plants in all crosses, indicating the widespread presence of non-allelic interactions. For number of leaf nodes at flowering the significant genetic effects on the generation means (analysed as described by Mather and Jinks, 1971; Rowe and Alexander, 1980) did not show a clear pattern (data not presented).   In the case of time to flowering the pattern was clearer (Table 1), the dominance effects being positive

in nearly all cases, and consistently larger in low-temperature treated plants than in control plants. Non-allelic interaction effects were more common and larger in cases where the dominance effect

Table 1.  Estimates of significant genetic effects for genetic models using six generation means for time to flowering

	Control				Low-temperature treated
Cross	Parameter^1	Estimate	SE	Parameter^1		Estimate	SE
TB14 ´ TB16	m	107.09	0.59	m		62.40	18.30
	[d]	12.61	1.85	[d]		17.66	1.00
				[h]		98.70	43.70
				[i]		32.20	18.30
				[l]		-61.80	26.10
c^2for model = 46.41  df = 1                                c^2for model = 555.19  df = 4
Residual c^2 = 4.35    df = 4   P = 0.360             Residual c^2 = 1.59     df = 1   P = 0.207
TB14 ´ TB28	m	91.15	0.29	m		26.85
	[d]	31.25	0.29	[d]		28.45
	[h]	-14.53	1.72	[h]		132.75
	[j]	-36.13	1.38	[i]		57.20
	[l]	41.88	2.20	[j]		-36.90
				[l]		-60.90
c^2for model = 763.39  df = 4                                c^2for model = 1813.00  df = 5
Residual c^2 =  0.04     df = 1   P = 0.123
TB14 ´ TB29	m	52.40		m		28.75
	[d]	6.60		[d]		2.35
	[h]	160.00		[h]		211.95
	[i]	63.40		[i]		81.40
	[j]	17.00		[j]		24.30
	[l]	-100.40		[l]		-136.10
c^2for model = 141.00  df = 5                               c^2for model = 121.60  df = 5
TB14 ´ TB33	m	110.22	1.09	m		21.78	6.37
	[d]	11.61	1.10	[d]		12.65	0.82
	[h]	13.64	1.72	[h]		198.60	16.80
				[i]		77.98	6.30
				[l]		-115.00	10.90
c^2for model = 481.87  df = 2                                c^2for model = 132.87  df = 4
Residual c^2 = 2.98      df = 3   P = 0.395               Residual c^2 = 0.29  df = 1   P = 0.590

¹Key:

Parameter	Gene effect	Parameter	Gene effect
m	mean	[i]	additive ´ additive
[d]	additive	[j]	additive ´ dominance
[h]	dominance	[l]	dominance ´ dominance

was large, as might be expected. In the case of stem height at flowering, dominance effects were variable in direction, and of about the same magnitude as the additive effects, except in the case of TB 14 ´ TB 28 where there was a substantial negative effect of dominance, especially in the low-temperature treated plants (data not presented).   The estimated narrow-sense heritability of number of leaf nodes at flowering in the control plants was zero in both crosses, but the corresponding values for the low-temperature treatment were moderately high (Table 2). Fairly high narrow-sense heritability was observed for time to flowering following both treatments in TB 14 ´ TB 28, but the corresponding values were low or zero in TB 14 ´ TB 33. For stem height in the control plants the narrow-sense heritability was fairly high in both crosses, but the corresponding values for the low-temperature treatment were low. The broad-sense heritabilities were sometimes higher than the narrow-sense heritabilities, as expected, but sometimes lower due to negative estimates of the component of variance for dominance. They generally followed the pattern of the narrow-sense heritabilities, the most notable exceptions being stem height in the low-temperature treated plants of TB 14 ´ TB 28, and time to flowering in control plants of TB 14 ´ TB 33, where the broad-sense heritabilities were substantially higher than the narrow sense.

Table 2.  Heritability estimates of characters measured in parent and progeny generations of crosses between inbred B. napus var annua genotypes

Treatment	Cross	Heritability	Character
			No. of leaf nodes	Time to flowering	Stem height
Control	TB 14 ´ TB 28	h^2N	0	0.79	0.85
	TB 14 ´ TB 33		0	0.17	0.64
Low temp.	TB 14 ´ TB 28	h^2N	0.68	0.56	0.32
	TB 14 ´ TB 33		0.37	0	0
Control	TB 14 ´ TB 28	h^2B	0	0.66	0.79
	TB 14 ´ TB 33		0.39	0.73	0.55
Low temp.	TB 14 ´ TB 28	h^2B	0.77	0.93	0.72
	TB 14 ´ TB 33		0	0	0

Molecular  analyses

The average time to flowering in non-thermoinduced F₁ plants was intermediate between that of B. napus (TB 14) and B. rapa cv. IB-5. Time to flowering in the BC₁  and BC₂  plants varied widely, but the mean values of successive backcrosses were progressively closer to that of the recurrent parent.   The   following   probe-enzyme  combinations  were  found  to  distinguish  the parents and the F₁ progeny: wg6b10/HindIII, wg7b3/HindIII, wg7f3/HindIII and wg8g1/EcoRV. All six F₁ plants showed the bands corresponding to both parents confirming that they were true hybrids.   The number of chromosomes in BC₁ plants ranged from 22 to 28 and that in BC₂ plants ranged from 22 to 26. This implies that BC₁ plants carry between 2 and 8 unpaired C genome chromosomes, whereas BC₂ plants carry between 2 and 6 unpaired C genome chromosomes, which should have been randomly distributed to gametes produced by F₁ plants.

The RFLP banding patterns of the five BC₂ plants of batch 2, BC₂ 1, BC₂ 2, BC₂ 5,   BC₂ 8 and BC₂ 10 with the probe wg7b3 showed that the plants BC₂ 1 and BC₂ 2 with early-flowering times (55 and 50 days respectively) have a similar pattern, whereas the plants BC₂ 5, BC₂ 8 and BC₂ 10 with the longest flowering times (92, 78 and 75 days respectively) possess a similar banding pattern, which is different to that of the early-flowering plants. The fragments 5.0kb and 10.5kb unique to B. napus are those associated with the late-flowering character. (The late-flowering plant BC₂ 4 of batch 2 (70 days) also had these two fragments.) This suggests that the late-flowering BC₂ plants could have the v₂ locus which determines the vernalization requirement for early flowering, on a C genome chromosome.

DISCUSSION

Though the patterns produced by the analysis of generation means were  complex, especially for the number of leaf nodes at flowering, it is clear that there are strong genetic influences on the vernalization response for all three characters in the genotypes studied here, and that the nature of these effects differs between the characters.   Diallel analysis of the same variables has also indicated that the response to vernalization could be manipulated in a breeding programme, both in screening cultivars to be used as parents and in screening progeny (S.R. Dahanayake, 1998). However, the large effects of dominance, and perhaps also non-allelic interactions, detected here indicate that selection in early-generations progeny may be misleading, particularly for time to flowering following a low-temperature stimulus.   The number of leaf nodes at flowering had fairly high narrow-sense heritability when the vernalization requirement was satisfied, but not under control conditions, indicating that this character will not be a good basis on which to select for the absence of a vernalization response.   In the absense of a low-temperature stimulus, time to flowering and stem height at flowering will be easier to influence by selection than number of leaf nodes.   The same pattern, i.e. higher narrow-sense heritability of number of leaf nodes following low-temperature treatment but of time to flowering and stem height in control plants, was also found in a diallel cross among similar genetic stocks (S.R. Dahanayake, 1998).  The heritability of the latter two characters was especially high in the cross TB 14 ´ TB 28, but even the narrow-sense heritability of 0.17 observed for time to flowering in TB 14 ´ TB 33 should be sufficient for selection. The high values obtained for broad-sense heritability of time to flowering and stem height at flowering in both crosses under non-vernalizing conditions suggest that for these two characters, considerable progress may be expected from visual selection for absence of a vernalization response in late generations of segregating populations, when non-allelic interactions are fixed.

Though southern hybridization using probe wg7b3 in the BC₂ population was able to detect segregation associated with late maturity, ascribable to a locus in the C genome, it was not possible to identify an individual chromosome associated with this polymorphism.   However, this should be possible if a greater number of BC₂ plants are screeneed. This would facilitate marker asssociated selection to eliminate the vernalization requirement for early flowering.

REFERENCES

Dahanayake, S. R. (1998) Genetic and Physiological Studies of the Vernalization Process in Spring Oilseed Rape (Brassica napus var. annua). P.D. thesis. The University of Western Australia.

Mather, K. and Jinks, J. L. (1971) Biometrical Genetics. London: Chapman and Hall.

Rowe, K. E. and Alexander, W. L. (1980). Computations for estimating the genetic parameters in joint-scaling tests. Crop Science 20: 109-110.

Thurling, N. and Vijendra Das, L. D. (1977) Variation in the pre-anthesis          development of spring rape (Brassica napus L.). Australian Journal of Agricultural Research 28, 597-607.

Thurling N., and Vijendra Das, L. D. (1979a). Genetic control of pre-anthesis development of spring rape (Brassica napus L.). I  Diallel analysis of variation in the field. Australian Journal of Agricultural Research 30, 251-59.

Thurling, N., and Vijendra Das, L. D. (1979b). Genetic control of pre-anthesis development of spring rape (Brassica napus L.). II Identification of individual genes controlling developmental pattern. Australian Journal of Agricultural Research 30, 261-71.